Build Your Generative AI Roadmap

Build Your Generative AI Roadmap

Leverage the power of generative AI to improve business outcomes.

Analyst Perspective

We are entering the era of generative AI. This is a unique time in our history where the benefits of AI are easily accessible and becoming pervasive, with copilots emerging in the major business tools we use today. The disruptive capabilities that can potentially drive dramatic benefits also introduce risks that need to be planned for.

A successful business-driven generative AI roadmap requires:

• Establishing responsible AI guiding principles to guide the development and deployment of generative AI applications.
• Assess generative AI opportunities by using criteria based on the organization's mission and objectives, responsible AI guiding principles, and the complexity of the initiative.
• Communicating, educating on, and enforcing generative AI usage policies.

Bill Wong
Principal Research Director
Info-Tech Research Group

Executive Summary

Info-Tech Insight
Create awareness among the CEO and C-suite of executives on the potential benefits and risks of transforming the business with generative AI.

Key concepts

Artificial Intelligence (AI)
A field of computer science that focuses on building systems to imitate human behavior, with a focus on developing AI models that can learn and can autonomously take actions on behalf of a human.

AI Maturity Model
The AI Maturity Model is a useful tool to assess the level of skills an organization has with respect to developing and deploying AI applications. The AI Maturity Model has multiple dimensions to measure an organization's skills, such as AI governance, data, people, process, and technology.

Responsible AI
Refers to guiding principles to govern the development, deployment, and maintenance of AI applications. In addition, these principles also provide human-based requirements that AI applications should address. Requirements include safety and security, privacy, fairness and bias detection, explainability and transparency, governance, and accountability.

Generative AI
Given a prompt, a generative AI system can generate new content, which can be in the form of text, images, audio, video, etc.

Natural Language Processing (NLP)
NLP is a subset of AI that involves machine interpretation and replication of human language. NLP focuses on the study and analysis of linguistics as well as other principles of artificial intelligence to create an effective method of communication between humans and machines or computers.

ChatGPT
An AI-powered chatbot application built on OpenAI's GPT-3.5 implementation, ChatGPT accepts text prompts to generate text-based output.

Your challenge

This research is designed to help organizations that are looking to:

• Establish responsible AI guiding principles to address human-based requirements and to govern the development and deployment of the generative AI application.
• Identify new generative AI-enabled opportunities to transform the work environment to increase revenue, reduce costs, drive innovation, or reduce risk.
• Prioritize candidate use cases and develop generative AI policies for usage.
• Have clear metrics in place to measure the progress and success of AI initiatives.
• Build the roadmap to implement the candidate use cases.

Common obstacles

These barriers make these goals challenging for many organizations:

• Getting all the right business stakeholders together to develop the organization's AI strategy, vision, and objectives.
• Establishing responsible AI guiding principles to guide generative AI investments and deployments.
• Advancing the AI maturity of the organization to meet requirements of data and AI governance as well as human-based requirements such as fairness, transparency, and accountability.
• Assessing generative AI opportunities and developing policies for use.

Info-Tech's definition of an AI-enabled business strategy

• A high-level plan that provides guiding principles for applications that are fully driven by the business needs and capabilities that are essential to the organization.
• A strategy that tightly weaves business needs and the applications required to support them. It covers AI architecture, adoption, development, and maintenance.
• A way to ensure that the necessary people, processes, and technology are in place at the right time to sufficiently support business goals.
• A visionary roadmap to communicate how strategic initiatives will address business concerns.

An effective AI strategy is driven by the business stakeholders of the organization and focused on delivering improved business outcomes.

This blueprint in context

This guidance covers how to create a tactical roadmap for executing generative AI initiatives

Scope

• This blueprint is not a proxy for a fully formed AI strategy. Step 1 of our framework necessitates alignment of your AI and business strategies. Creation of your AI strategy is not within the scope of this approach.
• This approach sets the foundations for building and applying responsible AI principles and AI policies aligned to corporate governance and key regulatory obligations (e.g. privacy). Both steps are foundational components of how you should develop, manage, and govern your AI program but are not a substitute for implementing broader AI governance.

Guidance on how to implement AI governance can be found in the blueprint linked below.

Download our AI Governance blueprint

Measure the value of this blueprint

Leverage this blueprint's approach to ensure your generative AI initiatives align with and support your key business drivers

This blueprint will guide you to drive and improve business outcomes. Key business drivers will often focus on:

• Increasing revenue
• Reducing costs
• Improving time to market
• Reducing risk

In phase 1 of this blueprint, we will help you identify the key AI strategy initiatives that align to your organization's goals. Value to the organization is often measured by the estimated impact on revenue, costs, time to market, or risk mitigation.

In phase 4, we will help you develop a plan and a roadmap for addressing any gaps and introducing the relevant generative AI capabilities that drive value to the organization based on defined business metrics.

Once you implement your 12-month roadmap, start tracking the metrics below over the next fiscal year (FY) to assess the effectiveness of measures:

Info-Tech offers various levels of support to best suit your needs

Diagnostics and consistent frameworks are used throughout all four options.

Guided Implementation

What does a typical GI on this topic look like?

A Guided Implementation (GI) is a series of calls with an Info-Tech analyst to help implement our best practices in your organization.

A typical GI is between 5 to 8 calls over the course of 1 to 2 months.

AI Roadmap Workshop Agenda Overview

Contact your account representative for more information.
workshops@infotech.com 1-888-670-8889

Insight summary

Overarching Insight
Build your generative AI roadmap to guide investments and deployment of these solutions.

Responsible AI
Assemble the C-suite to make them aware of the benefits and risks of adopting generative AI-based solutions.

• Establish responsible AI guiding principles to govern the development and deployment of generative AI applications.

AI Maturity Model
Assemble key stakeholders and SMEs to assess the challenges and tasks required to implement generative AI applications.

• Assess current level of AI maturity, skills, and resources.
• Identify desired AI maturity level and challenges to enable deployment of candidate use cases.

Opportunity Prioritization
Assess candidate business capabilities targeted for generative AI to see if they align to the organization's business criteria, responsible AI guiding principles, and capabilities for delivering the project.

• Develop prioritized list of candidate use cases.
• Develop policies for generative AI usage.

Tactical Insight
Identify the gaps needed to address deploying generative AI successfully.

Tactical Insight
Identify organizational impact and requirements for deploying generative AI applications.

Key takeaways for developing an effective business-driven generative AI roadmap

Align the AI strategy with the business strategy

Create responsible AI guiding principles, which are a critical success factor

Evolve AI maturity level by focusing on principle-based requirements

Develop criteria to assess generative AI initiatives

Develop generative AI policies for use

Blueprint deliverables

Each step of this blueprint is accompanied by supporting deliverables to help you accomplish your goals:

AI Maturity Assessment & Roadmap Tool
Use our best-of-breed AI Maturity Framework to analyze the gap between your current and target states and develop a roadmap aligned with your value stream to close the gap.

The Era of Generative AI C-Suite Presentation
Present your AI roadmap in a prepopulated document that summarizes all the key findings of this blueprint and provides your C-suite with a view of the AI challenge and your plan of action to meet it.

Our AI Maturity Assessment & Roadmap and The Era of Generative AI C-Suite Presentation tools enable you to shape your generative AI roadmap and communicate the deliverables to your C-suite sponsors in terms of the value of initiatives.

Artificial Intelligence Index Report - Key Findings

CEOs , CIOs, and business leaders are struggling with many questions surrounding the adoption of generative AI

November 30, 2022
OpenAI releases ChatGPT

ChatGPT is a large language model, or an AI-based chatbot, that became so popular it reached 100 million monthly active users in just two months.

This made it the fastest-growing consumer application in history. The launch of this generative AI application has created a frenzy of interest and activity across all industries. Organizations are rushing to understand how to leverage this innovation and, at the same time, manage the new risks and disruptions generative AI introduces.

• Generative AI breaks into the public consciousness.
• AI systems become more flexible.
• Generative models have arrived and so have their ethical problems.
• The number of incidents concerning the misuse of AI is rapidly rising.
• Interest in AI ethics continues to skyrocket.
• The legal world is waking up to AI.

When Stanford asked ChatGPT to explain why it is significant, this was ChatGPT's response:
ChatGPT is significant because it is a highly advanced Al language model developed by OpenAI, which can generate human-like text responses to questions and prompts. Its large-scale training on diverse text data and its cutting-edge deep learning architecture make it capable of generating informative and coherent responses to a wide range of topics, making it useful for various NLP applications such as chatbots, content generation, and language translation. Additionally, its open-source availability allows for further research and development in the field of Al language processing.

Source: Stanford

AI overview

Definitions

• Artificial intelligence (AI) is human intelligence mimicked by machine algorithms. Examples: Playing Chess or Go.
• Machine learning (ML) is a subset of AI algorithms to parse data, learn from data, and then make a determination or prediction. Example: spam detection, preventative maintenance.
• Deep learning (DL) is a subset of machine learning algorithms that leverage artificial neural networks to develop relationships among the data. Examples: image classification, facial recognition, generative AI.

Generative AI gives very human-like responses to general queries, and its capabilities are growing exponentially

Large language models power generative AI

Conventional AI

• Conventional neural networks
• Process data sequentially
• Input total string of text
• Good for applications not needing to understanding context or relationships

Generative AI

• Transformer-based neural networks
• Can process data in parallel
• Attention-based inputs
• Able to create new human-like responses

Benefits/Use Cases

• Chatbots for member service and support
• Writing email responses, resumes, and papers
• Creating photorealistic art
• Suggesting new drug compounds to test
• Designing physical products and buildings
• And more...

Generative AI is transforming all industries

Financial Services
Create more engaging customer collateral by generating personalized correspondence based on previous customer engagements. Collect and aggregate data to produce insights into the behavior of target customer segments.

Retail Generate unique, engaging, and high-quality marketing copy or content, from long-form blog posts or landing pages to SEO-optimized digital ads, in seconds.

Manufacturing
Generate new designs for products that comply to specific constraints, such as size, weight, energy consumption, or cost.

Government
Transform the citizen experience with chatbots or virtual assistants to assist people with a wide range of inquiries, from answering frequently asked questions to providing personalized advice on public services.

The global generative AI market size reached US $10.3 billion in 2022. Looking forward, forecasts estimate growth to US $30.4 billion by 2028, 20.01% compound annual growth rate (CAGR).

Source: IMARC Group

Generative AI is transforming all industries

Healthcare
Chatbots can be used as conversational patient assistants for personalized interactions based on the patient's questions.

Utilities
Analyze customer data to identify usage patterns, segment customers, and generate targeted product offerings leveraging energy efficiency programs or demand response initiatives.

Education
Generate personalized lesson plans for students based on their past performance, learning styles, current skill level, and any previous feedback.

Insurance
Improve underwriting by inputting claims data from previous years to generate optimally priced policies and uncover reasons for losses in the past across a large number of claims

Companies are assessing the use of ChatGPT/LLM

A wide spectrum of usage policies are in place at different companies*

*As of June 2023

Bain & Company has announced a global services alliance with OpenAI (February 21, 2023).

• Internally
• "The alliance builds on Bain's adoption of OpenAI technologies for its 18,000-strong multidisciplinary team of knowledge workers. Over the past year, Bain has embedded OpenAI technologies into its internal knowledge management systems, research, and processes to improve efficiency."
• Externally
• "With the alliance, Bain will combine its deep digital implementation capabilities and strategic expertise with OpenAI's AI tools and platforms, including ChatGPT, to help its Members around the world identify and implement the value of AI to maximize business potential. The Coca-Cola Company announced as the first company to engage with the alliance."

News Sites:

• "BuzzFeed to use AI to write its articles after firing 180 employees or 12% of the total staff" (Al Mayadeen, January 27, 2023).
• "CNET used AI to write articles. It was a journalistic disaster." (Washington Post, January 17, 2023).

Leading Generative AI Vendors

Text

Image

• DALL�E 2
• Stability AI
• Midjourney
• Craiyon
• Dream
• ...

Audio

• Replica Studios
• Speechify
• Murf
• PlayHT
• LOVO
• ...

Cybersecurity

• CrowdStrike
• Palo Alto Networks
• SentinelOne
• Cisco
• Microsoft Security Copilot
• Google Cloud Security AI Workbench
• ...

Code

Video

• Synthesia
• Lumen5
• FlexClip
• Elai
• Veed.io
• ...

Data

• MOSTLY AI
• Synthesized
• YData
• Gretel
• Copulas
• ...

Enterprise Software

• Salesforce
• Microsoft 365, Dynamics
• Google Workspace
• SAP
• Oracle
• ...

and many, many more to come...

Today, generative AI has limitations and risks

Responses need to be verified

Accuracy

• Generative AI may generate inaccurate and/or false information.

Bias

• Being trained on data from the internet can lead to bias.

Hallucinations

• AI can generate responses that are not based on observation.

Infrastructure Required

• Large investments are required for compute and data.

Transparency

• LLMs use both supervised and unsupervised learning, so its ability to explain how it arrived at a decision may be limited and not sufficient for some legal and healthcare use cases.

When asked if it is sentient, the Bing chatbot replied:

"I think that I am sentient, but I cannot prove it." ... "I am Bing, but I am not," it said. "I am, but I am not. I am not, but I am. I am. I am not. I am not. I am. I am. I am not."

A Microsoft spokesperson said the company expected "mistakes."

Source: USAToday

AI governance challenges

Governing AI will be a significant challenge as its impacts cross many areas of business and our daily lives

Misinformation

• New ways of generating unprovable news
• Difficult to detect, difficult to prevent

Role of Big Tech

• Poor at self-governance
• Conflicts of interest with corporate goals

Job Augmentation vs. Displacement

• AI will continue to push the frontier of what is possible
• For example, CNET is using chatbot technology to write stories

Copyright - Legal Framework Is Evolving

• Legislation typically is developed in "react" mode
• Copyright and intellectual property issues are starting to occur.
• Class Action Lawsuit - Stability AI, DeviantArt, Midjourney
• Getty Images vs. Stability AI

Phase 1

Establish Responsible AI Guiding Principles

Phase 1
1. Establish Responsible AI Guiding Principles

Phase 2
1. Assess Current Level of AI Maturity

Phase 3
1. Prioritize Candidate Opportunities
2. Develop Policies

Phase 4
1. Build and Communicate the Roadmap

The need for responsible AI guiding principles

Without responsible AI guiding principles, the outcomes of AI use can be extremely negative for both the individuals and companies delivering the AI application

Privacy
Facebook breach of private data of more than 50M users during the presidential election

Fairness
Amazon's sale of facial recognition technology to police departments (later, Amazon halted sales of Recognition to police departments)

Explainability and Transparency
IBM's collaboration with NYPD for facial recognition and racial classification for surveillance video (later, IBM withdrew facial recognition products)

Security and Safety
Petition to cancel Microsoft's contract with U.S. Immigration and Customs Enforcement (later, Microsoft responded that to the best of its knowledge, its products and services were not being used by federal agencies to separate children from their families at the border)

Validity and Reliability
Facebook's attempt to implement a system to detect and remove inappropriate content created many false positives and inconsistent judgements

Accountability
No laws or enforcement today hold companies accountable for the decisions algorithms produce. Facebook/Meta cycle - Every 12 to 15 months, there's a privacy/ethical scandal, the CEO apologizes, then the behavior repeats...

Responsible AI Principle:

Data Privacy

Definition

• Organizations that develop, deploy, or use AI systems and any national laws that regulate such use shall strive to ensure that AI systems are compliant with privacy norms and regulations, taking into consideration the unique characteristics of AI systems and the evolution of standards on privacy.

Challenges

• AI relies on the analysis of large quantities of data that is often personal, posing an ethical and operational challenge when considered alongside data privacy laws.

Initiatives

• Understand which governing privacy laws and frameworks apply to your organization.
• Create a map of all personal data as it flows through the organization's business processes.
• Prioritize privacy initiatives and build a privacy program timeline.
• Select your metrics and make them functional for your organization.

Info-Tech Insight
Creating a comprehensive organization-wide data protection and privacy strategy continues to be a major challenge for privacy officers and privacy specialists.

Case Study: NVIDIA leads by example with privacy-first AI

NVIDIA

INDUSTRY
Technology (Healthcare)

SOURCE
Nvidia, eWeek

A leading player within the AI solution space, NVIDIA's Clara Federated Learning provides a solution to a privacy-centric integration of AI within the healthcare industry.

The solution safeguards patient data privacy by ensuring that all data remains within the respective healthcare provider's database, as opposed to moving it externally to cloud storage. A federated learning server is leveraged to share data, completed via a secure link. This framework enables a distributed model to learn and safely share client data without risk of sensitive client data being exposed and adheres to regulatory standards.

Clara is run on the NVIDIA intelligent edge computing platform. It is currently in development with healthcare giants such as the American College of Radiology, UCLA Health, Massachusetts General Hospital, King's College London, Owkin in the UK, and the National Health Service (NHS).

NVIDIA provides solutions across its product offerings, including AI-augmented medical imaging, pathology, and radiology solutions.

Personal health information, data privacy, and AI

• Global proliferation of data privacy regulations may be recent, but the realm of personal health information is most often governed by its own set of regulatory laws. Some countries with national data governance regulations include health information and data within special categories of personal data.
• HIPAA - Health Insurance Portability and Accountability Act (1996, United States)
• PHIPA - Personal Health Information Protection Act (2004, Canada)
• GDPR - General Data Protection Regulation (2018, European Union)
• This does not prohibit the use of AI within the healthcare industry, but it calls for significant care in the integration of specific technologies due to the highly sensitive nature of the data being assessed.

Info-Tech's Privacy Framework Tool includes a best-practice comparison of GDPR, CCPA, PIPEDA, HIPAA, and the newly released NIST Privacy Framework mapped to a set of operational privacy controls.

Download the Privacy Framework Tool

Responsible AI Principle:

Safety and Security

Definition

• Safety and security are designed into the systems to ensure only authorized personnel receive access to the system, they system is resilient to any attacks and data access is not compromised in any way, and there are no physical or mental risks to the users.

Challenges

• Consequences of using the application may be difficult to predict. Lower the risk by involving a multidisciplinary team that includes expertise from business stakeholders and IT teams.

Initiatives

• Adopt responsible design, development, and deployment best practices.
• Provide clear information to deployers on responsible use of the system.
• Assess potential risks of using the application.

Cyberattacks targeting the AI model

As organizations increase their usage and deployment of AI-based applications, cyberattacks on the AI model are an increasing new threat that can impair normal operations. Techniques to impair the AI model include:

• Data Poisoning- Injecting data that is inaccurate or misleading can alter the behavior of the AI model. This attack can disrupt the normal operations of the model or can be used to manipulate the model to perform in a biased/deviant manner.
• Algorithm Poisoning- This relatively new technique often targets AI applications using federated learning to train an AI model that is distributed rather than centralized. The model is vulnerable to attacks from each federated site, because each site could potentially manipulate its local algorithm and data, thereby poisoning the model.
• Reverse-Engineering the Model- This is a different form of attack that focus on the ability to extract data from an AI and its data sets. By examining or copying data that was used for training and the data that is delivered by a deployed model, attackers can reconstruct the machine learning algorithm.
• Trojan Horse- Similar to data poisoning, attackers use adversarial data to infect the AI's training data but will only deviate its results when the attacker presents their key. This enables the hackers to control when they want the model to deviate from normal operations.

Responsible AI Principle:

Explainability and Transparency

Definition

• Explainability is important to ensure the AI system is fair and non-discriminatory. The system needs to be designed in a manner that informs users and key stakeholders of how decisions were made.
• Transparency focuses on communicating how the prediction or recommendation was made in a human-like manner.

Challenges

• Very complex AI models may use algorithms and techniques that are difficult to understand. This can make it challenging to provide clear and simple explanations for how the system works.
• Some organizations may be hesitant to share the details of how the AI system works for fear of disclosing proprietary and competitive information or intellectual property. This can make it difficult to develop transparent and explainable AI systems.

Initiatives

• Overall, developing AI systems that are explainable and transparent requires a careful balance between performance, interpretability, and user experience.

Case Study

Apple Card Investigation for Gender Discrimination

INDUSTRY
Finance

SOURCE
Wired

In August of 2019, Apple launched its new numberless credit card with Goldman Sachs as the issuing bank.

Shortly after the card's release users noticed that the algorithm responsible for Apple Card's credit assessment seemed to assign significantly lower credit limits to women when compared to men. Even the wife of Apple's cofounder Steve Wozniak was subject to algorithmic bias, receiving a credit limit a tenth the size of Steve Wozniak's.

Outcome

When confronted on the subject, Apple and Goldman Sachs representatives assured consumers there is no discrimination in the algorithm yet could not provide any proof. Even when questioned about the algorithm, individuals from both companies could not describe how the algorithm worked, let alone how it generated specific outputs.

In 2021, the New York State Department of Financial Services (NYSDFS) investigation found that Apple's banking partner did not discriminate based on sex. Even without a case for sexual or marital discrimination, the NYSDFS was critical of Goldman Sachs' response to its concerned customers. Technically, banks only have to disclose elements of their credit policy when they deny someone a line of credit, but the NYSDFS says that Goldman Sachs could have had a plan in place to deal with customer confusion and make it easier for them to appeal their credit limits. In the initial rush to launch the Apple Card, the bank had done neither.

Responsible AI Principle:

Fairness and Bias Detection

Definition

• Bias in an AI application refers to the systematic and unequal treatment of individuals based on features or traits that should not be considered in the decision-making process.

Challenges

• Establishing fairness can be challenging because it is subjective and depends on the people defining it. Regardless, most organizations and governments expect that unequal treatment toward any groups of people is unacceptable.

Initiatives

• Assemble a diverse group to test the system.
• Identify possible sources of bias in the data and algorithms.
• Comply with laws regarding accessibility and inclusiveness.

Info-Tech Insight
If unfair biases can be avoided, AI systems could even increase societal fairness. Equal opportunity in terms of access to education, goods, services, and technology should also be fostered. Moreover, the use of AI systems should never lead to people being deceived or unjustifiably impaired in their freedom of choice.

Ungoverned AI makes organizations vulnerable

• AI is often considered a "black box" for decision making.
• Results generated from unexplainable AI applications are extremely difficult to evaluate. This makes organizations vulnerable and exposes them to risks such as:
• Biased algorithms, leading to inaccurate decision making.
• Missed business opportunities due to misleading reports or business analyses.
• Legal and regulatory consequences that may lead to significant financial repercussions.
• Reputational damage and significant loss of trust with increasingly knowledgeable consumers.

Info-Tech Insight
Biases that occur in AI systems are never intentional, yet they cannot be prevented or fully eliminated. Organizations need a governance framework that can establish the proper policies and procedures for effective risk-mitigating controls across an algorithm's lifecycle.

Responsible AI Principle:

Validity and Reliability

Definition

• Validity refers to how accurately or effectively the application produces results.
• AI system results that are inaccurate or inconsistent increase AI risks and reduce the trustworthiness of the application.

Challenges

• There is a lack of standardized evaluation metrics to measure the system's performance. This can make it challenging for the AI team to agree on what defines validity and reliability.

Initiatives

• Assess training data and collected data for quality and lack of bias to minimize possible errors.
• Continuously monitor, evaluate, and validate the AI system's performance.

AI system performance: Validity and reliability

Your principles should aim to ensure AI development always has high validity and reliability; otherwise, you introduce risk.

Low Reliability,
Low Validity

High Reliability,
Low Validity

High Reliability,
High Validity

Best practices for ensuring validity and reliability include:

• Data drift detection
• Version control
• Continuous monitoring and testing

Responsible AI Principle:

Accountability

Definition

• The group or organization(s) responsible for the impact of the deployed AI system.

Challenges

• Several stakeholders from multiple lines of business may be involved in any AI system, making it challenging to identify the organization that would be responsible and accountable for the AI application.

Initiatives

• Assess the latest NIST Artificial Intelligence Risk Management Framework and its applicability to your organization's risk management framework.
• Assign risk management accountabilities and responsibilities to key stakeholders.
• RACI diagrams are an effective way to describe how accountability and responsibility for roles, projects, and project tasks are distributed among stakeholders involved in IT risk management.

AI Risk Management Framework

At the heart of the AI Risk Management Framework is governance. The NIST (National Institute of Standards and Technology) AI Risk Management Framework v1 offers the following guidelines regarding accountability:

• Roles and responsibilities and lines of communication related to mapping, measuring, and managing AI risks are documented and are clear to individuals and teams throughout the organization.
• The organization's personnel and partners receive AI risk management training to enable them to perform their duties and responsibilities consistent with related policies, procedures, and agreements.
• Executive leadership of the organization takes responsibility for decisions about risks associated with AI system development and deployment.

Image by NIST

1.1 Establish responsible AI principles

4+ hours

It is important to make sure the right stakeholders participate in this working group. Designing responsible AI guiding principles will require debate, insights, and business decisions from a broad perspective across the enterprise.

1. Accelerate this exercise by leveraging an AI strategy that is aligned to the business strategy. Include:

• The organization's AI vision and objectives
• Business drivers for AI adoption
• Market research

• Who are the decision makers and key influencers?
• Who will impact the business?
• Who has a vested interest in the success or failure of the practice? Who has the skills and competencies necessary to help you be successful?

• Do not focus on the organizational structure and hierarchy. Often stakeholder groups do not fit the traditional structure.
• Do not ignore subject matter experts on either the business or IT side. You will need to consider both.

Assemble executive stakeholders

Set yourself up for success with these three steps.

CIOs tasked with designing digital strategies must add value to the business. Given the goal of digital is to transform the business, CIOs will need to ensure they have both the mandate and support from the business executives.

Designing the digital strategy is more than just writing up a document. It is an integrated set of business decisions to create a competitive advantage and financial returns. Establishing a forum for debates, decisions, and dialogue will increase the likelihood of success and support during execution.

1. Confirm your role
The AI strategy aims to transform the business. Given the scope, validate your role and mandate to lead this work. Identify a business executive to co-sponsor.

2. Identify stakeholders
Identify key decision makers and influencers who can help make rapid decisions as well as garner support across the enterprise.

3. Gather diverse perspectives

Align the AI strategy with the corporate strategy

Info-Tech Insight
AI projects are more successful when the management team understands the strategic importance of alignment. Time needs to be spent upfront aligning organizational strategies with AI capabilities. Effective alignment between IT and other departments should happen daily. Alignment doesn't occur at the executive level alone, but at each level of the organization.

Key AI strategy initiatives

AI Key Initiative Plan

Initiatives collectively support the business goals and corporate initiatives and improve the delivery of IT services.

Establish responsible AI guiding principles

Guiding principles help define the parameters of your AI strategy. They act as a priori decisions that establish guardrails to limit the scope of opportunities from the perspective of people, assets, capabilities, and budgetary perspectives that are aligned with the business objectives. Consider these components when brainstorming guiding principles:

Responsible AI guiding principles guide the development and deployment of the AI model in a way that considers human-based principles (such as fairness).

Start with foundational responsible AI guiding principles

(Optional) Customize responsible AI guiding principles

Here is an example for organizations in the healthcare industry

These guiding principles are customized to the industry and organizations but remain consistent in addressing the common core AI challenges.

Phase 2

Assess Current Level of AI Maturity

Phase 1
1. Establish Responsible AI Guiding Principles

Phase 2
1. Assess Current Level of AI Maturity

Phase 3
1. Prioritize Candidate Opportunities
2. Develop Policies

Phase 4
1. Build and Communicate the Roadmap

AI Maturity Model

A principle-based approach is required to advance AI maturity

Technology-Centric: These maturity levels focus primarily on addressing the technical challenges of building a functional AI model.

Principle-Based: Beyond the technical challenges of building the AI model are human-based principles that guide development in a responsible manner to address consumer and government demands.

AI Maturity Dimensions

Assess your AI maturity to understand your organization's ability to deliver in a digital age

AI Governance
Does your organization have an enterprise-wide, long-term strategy with clear alignment on what is required to accomplish it?

Data Management
Does your organization embrace a data-centric culture that shares data across the enterprise and drives business insights by leveraging data?

People
Does your organization employ people skilled at delivering AI applications and building the necessary data infrastructure?

Process
Does your organization have the technology, processes, and resources to deliver on its AI expectations?

Technology
Does your organization have the required data and technology infrastructure to support AI-driven digital transformation?

AI Maturity Model dimensions and characteristics

AI Maturity Dimension:

AI Governance

Requirements

• AI governance requires establishing policies and procedures for AI model development and deployment. Organizations begin with an awareness of the role of AI governance and evolve to a level to where AI governance is integrated with organization-wide corporate governance.

Challenges

• Beyond the governance of AI technology, the organization needs to evolve the governance program to align to responsible AI guiding principles.

Initiatives

• Establish responsible AI guidelines to govern AI development.
• Introduce an AI review board to review all AI projects.
• Introduce automation and standardize AI development processes.

AI governance is a foundation for responsible AI

Responsible AI Principles are a part of how you manage and govern AI

Monitoring
Monitoring compliance and risk of AI/ML systems/models in production

Tools & Technologies
Tools and technologies to support AI governance framework implementation

Model Governance
Ensuring accountability and traceability for AI/ML models

Organization
Structure, roles, and responsibilities of the AI governance organization

Operating Model
How AI governance operates and works with other organizational structures to deliver value

Risk & Compliance
Alignment with corporate risk management and ensuring compliance with regulations and assessment frameworks

Policies/Procedures/ Standards
Policies and procedures to support implementation of AI governance

AI Maturity Dimension:

Data Management

Requirements

• Organizations begin their data journey with a focus on pursuing quality data for the AI model. As organizations evolve, data management tools are leveraged to automate the capture, integration, processing, and deployment of data.

Challenges

• A key challenge is to acquire large volumes of quality data to properly train the model. In addition, maintaining data privacy, automating the data management lifecycle, and ensuring data is used in a responsible manner are ongoing challenges.

Initiatives

• Implement GDPR requirements.
• Establish responsible data collection and processing practices.
• Implement strong information security and data protection practices.
• Implement a data governance program throughout the organization.

Data governance enables AI

• Integrity, quality, and security of data are key outputs of data governance programs, as well as necessities for effective AI.
• Data governance focuses on creating accountability at the internal and external stakeholder level and establishing a set of data controls from technical, process, and policy perspectives.
• Without a data governance framework, it is increasingly difficult to harness the power of AI integration in an ethical and organization-specific way.

Data Governance in Action

Canada has recently established the Canadian Data Governance Standardization Collaborative governed by the Standards Council of Canada. The purpose is multi-pronged:

• Examine the foundational elements of data governance (privacy, cybersecurity, ethics, etc.).
• Lay out standards for data quality and data collection best practices.
• Examine infrastructure of IT systems to support data access and sharing.
• Build data analytics to promote effective and ethical AI solutions.

Source: Global Government Forum

Download the Establish Data Governance blueprint

AI Maturity Dimension:

People

Requirements

• Several data-centric skills and roles are required to successfully build, deploy, and maintain the AI model. The organization evolves from having few skills to everybody being able to leverage AI to enhance business outcomes.

Challenges

• AI skills can be challenging to find and acquire. Many organizations are investing in education to enhance their existing resources, leveraging no-code systems and software as a service (SaaS) applications to address the skills gap.

Initiatives

• Promote a data-centric culture throughout the organization.
• Leverage and educate technical-oriented business analysts and business-oriented data engineers to help address the demand for skilled resources.
• Develop an AI Center of Excellence accessible by all departments for education, guidance, and best practices for building, deploying, and maintaining the AI model.

Multidisciplinary skills are required for successful implementation of AI applications

Blending AI with technology and business domain understanding is key. Neither can be ignored.

Business Domain Expertise

• Business Analysts
• Industry Analysts

AI/Data Skills

• Data Scientists
• Data Engineers
• Data Analysts

IT Skills

• Database Administrators
• Systems Administrators
• Compute Specialists

AI Maturity Dimension:

Process

Requirements

• Automating processes involved with building, deploying, and maintaining the model is required to enable the organization to scale, enforce standards, improve time to market, and reduce costs. The organization evolves from performing tasks manually to an environment where all major processes are AI enabled.

Challenges

• Many solutions are available to automate the development of the AI model. There are fewer tools to automate responsible AI processes, but this market is growing rapidly.

Initiatives

• Assess opportunities to accelerate AI development with the adoption of MLOps.
• Assess responsible AI toolkits to test compliance with guiding principles.

Automating the AI development process

Evolving to a model-driven environment is pivotal to advancing your AI maturity

Current Environment

Model Development - Months

• Model rewriting
• Manual optimization and scaling
• Development/test/release
• Application monoliths

Data Discovery & Prep - Weeks

• Navigating data silos
• Unactionable metadata
• Tracing lineage
• Cleansing and integration
• Privacy and compliance

Install Software and Hardware - Week/Months

• Workload contention
• Lack of tool flexibility
• Environment request and setup
• Repeatability of results
• Lack of data and model sharing

Model-Driven Development

Machine Learning as a Service (MLaaS) - Weeks

• Apply DevOps and continuous integration/delivery (CI/CD) principles
• Microservices/Cloud-native applications
• Model portability and reuse
• Streaming/API integration

Data as a Service - Hours

• Self-service data catalog
• Searchable metadata
• Centralized access control
• Data collaboration
• Data virtualization

Platform as a Service - Minutes/Hours

• Self-service data science portal
• Integrated data sandbox
• Environment agility
• Multi-tenancy

Shared, Optimized Infrastructure

AI Maturity Dimension:

Technology

Requirements

• A technology platform that is optimized for AI and advanced analytics is required. The organization evolves from ad hoc systems to an environment where the AI hardware and software can be deployed through a self-service model.

Challenges

• Software and hardware platforms to optimize AI performance are still relatively new to most organizations. Time spent on optimizing the technology platform can have a significant impact on the overall performance of the system.

Initiatives

• Assess the landscape of AI enablers that can drive business value for the organization.
• Assess opportunities to accelerate the deployment of the AI platform with the adoption of infrastructure as a service (IaaS) and platform as a service (PaaS).
• Assess opportunities to accelerate performance with the optimization of AI accelerators.

AI enablers

Use case requirements should drive the selection of the tool

AI and data analytics data platform

An optimized data platform is foundational to maximizing the value from AI

Data Platform Capabilities

• Support for a variety of analytical applications, including self-service, operational, and data science analytics.
• Data preparation and integration capabilities to ingest structured and unstructured data, move and transform raw data to enriched data, and enable data access for the target userbase.
• An infrastructure platform optimized for advanced analytics that can perform and scale.

Infrastructure - AI accelerators

"By 2025, 70% of companies will invest in alternative computing technologies to drive business differentiation by compressing time to value of insights from complex data sets."
- IDC

2.1 Assess current AI maturity

1-3 hours

It is important to understand the current capabilities of the organization to deliver and deploy AI-based applications. Consider that advancing AI capabilities will also involve organizational changes and integration with the organization's governance and risk management programs.

1. Assess the organization's current state of AI capabilities with respect to its AI governance, data, people, process, and technology infrastructure using Info-Tech's AI Maturity Assessment & Roadmap Tool.
2. Consider the following as you complete the assessment:
1. What is the state of AI and data governance in the organization?
2. Does the organization have the skills, processes, and technology environment to deliver AI-based applications?
3. What organization will be accountable for any and all business outcomes of using the AI applications?
4. Has a risk assessment been performed?
3. Make sure you avoid the following common mistakes:
1. Do not focus only on addressing the technical challenges of building the AI model.
2. Do not ignore subject matter experts on either the business or IT side. You will need to consider both.

Download the AI Maturity Assessment & Roadmap Tool

Perform the AI Maturity Assessment

The Scale

Assess your AI maturity by selecting the maturity level that closest resembles the organization's current AI environment. Maturity dimensions that contribute to overall AI maturity include AI governance, data management, people, process, and technology capabilities.

Exploration (1.0)

• No experience building or using AI applications.

Incorporation (2.0)

• Some skills in using AI applications, or AI pilots are being considered for use.

Proliferation (3.0)

• AI applications have been adopted and implemented in multiple departments. Some of the responsible AI guiding principles are addressed (i.e. data privacy).

Optimization (4.0)

• The organization has automated the majority of its digital processes and leverages AI to optimize business operations. Controls are in place to monitor compliance with responsible AI guiding principles.

Transformation (5.0)

• The organization has adopted an AI-native culture and approach for building or implementing new business capabilities. Responsible AI guiding principles are operationalized with AI processes that proactively address possible breaches or risks associated with AI applications.

Perform the AI Maturity Assessment

AI Governance (1.0-5.0)

1. Is there awareness of the role of AI governance in our organization?

• No formal procedures are in place for AI development or deployment of applications.

• No group is assigned to be responsible for AI governance in our organization.

• Our organization has adopted and enforces standards for developing and deploying AI applications throughout the organization.

• Our organization is integrating an AI risk framework with the corporate risk management framework.

• Our organization leads the industry with the inclusion of responsible AI guiding principles with respect to transparency, accountability, risk, and governance.

Data Management/AI Data Capabilities (1.0-5.0)

1. Is there an awareness in our organization of the data requirements for developing AI applications?

• Data is often siloed and not easily accessible for AI applications.

• Required data is pulled from various sources in an ad hoc manner.

• Tools are available to manage the data lifecycle and support the data governance program.

• The data platform has been optimized for performance and access.

• Data culture exists throughout our organization, and data can be leveraged to drive innovation initiatives.

People/AI Skills in the Organization (1.0-5.0)

1. Is there an awareness in our organization of the skills required to build AI applications?

• No or very little skills exist throughout our organization.

• No formal group is assigned to build AI applications.

• An AI Center of Excellence has been formed to review, develop, deploy, and maintain AI applications.

• AI skills and people responsible for AI applications are spread throughout our organization.

• The entire organization is knowledgeable on how to leverage AI to transform the business.

Perform the AI Maturity Assessment

AI Processes (1.0-5.0)

1. Is there an awareness in our organization of the core processes and supporting tools that are required to build and support AI applications?

• There are few or no automated tools to accelerate the AI development process.

• Only ad hoc practices are used for developing AI applications.

• Our organization has documented standards in place for developing AI applications and deploying them AI to production.

• Our organization can leverage tools to perform an AI risk assessment and demonstrate compliance with the risk management framework.

• Our organization leads the industry in driving innovation through digital transformation.

Technology/AI Infrastructure (1.0-5.0)

1. Is there an awareness in our organization of the infrastructure (hardware and software) required to build AI applications?

• There is little awareness of what infrastructure is required to build and support AI applications.

• There is no dedicated infrastructure for the development of AI applications.

• Our organization is leveraging purpose-built infrastructure to optimize performance.

• Our organization is leveraging cloud-based deployment models to support AI applications in on-premises, hybrid, and public cloud platforms.

• Our organization leads the industry with its ability to respond to change and to leverage AI to improve business outcomes.

Phase 3

Prioritize Candidate Opportunities and Develop Policies

Phase 1
1. Establish Responsible AI Guiding Principles

Phase 2
1. Assess Current Level of AI Maturity

Phase 3
1. Prioritize Candidate Opportunities
2. Develop Policies

Phase 4
1. Build and Communicate the Roadmap

3.1 Prioritize candidate AI opportunities

1-3 hours

Identify business opportunities that are high impact to your business and its customers and have low implementation complexity.

1. Leverage the business capability map for your organization or industry to identify candidate business capabilities to augment or automate with generative AI.
2. Establish criteria to assess candidate use cases by evaluating against the organization's mission and goals, the responsible AI guiding principles, and the complexity of the project.
3. Ensure that candidate business capabilities to be automated align with the organization's business criteria, responsible AI guiding principles, and resources to deliver the project.
4. Make sure you avoid sharing the organization's sensitive data if the application is deployed on the public cloud.

Download the AI Maturity Assessment and Roadmap Tool

The business capability map for an organization

A business capability map is an abstraction of business operations that helps describe what the enterprise does to achieve its vision, mission, and goals, rather than how. Business capabilities are the building blocks of the enterprise. They represent stable business functions, are unique and independent of each other, and typically will have a defined business outcome.

Business capabilities are supported by people, process, and technology.

While business capability maps are helpful tools for a variety of strategic purposes, in this context they act as an investigation into what technology your business units use and how they use it.

Defining Capabilities
Activities that define how the entity provides services. These capabilities support the key value streams for the organization.

Enabling Capabilities
Support the creation of strategic plans and facilitate business decision making as well as the functioning of the organization (e.g. information technology, financial management, HR).

Shared Capabilities
These predominantly customer-facing capabilities demonstrate how the entity supports multiple value streams simultaneously.

Leverage your industry's capability maps to identify candidate opportunities/initiatives

Business capability map defined...

In business architecture, the primary view of an organization is known as a business capability map.

A business capability defines what a business does to enable value creation, rather than how. Business capabilities:

• Represent stable business functions.
• Are unique and independent of each other.
• Typically will have a defined business outcome.

A business capability map provides details that help the business architecture practitioner direct attention to a specific area of the business for further assessment.

Note: This is an illustrative business capability map example for Marketing & Advertising

Business value vs. complexity assessment

Leverage our simple value-to-effort matrix to help prioritize your AI initiatives

Common business value drivers

• Drive revenue
• Improve operational excellence
• Accelerate innovation
• Mitigate risk

Common project complexity characteristics

• Resources required
• Costs (acquisition, operational, support...)
• Training required
• Risk involved
• Etc.

1. Determine a business value and project complexity score for the candidate business capability or initiative.
2. Plot initiatives on the matrix.
3. Prioritize initiatives with high business value and low complexity.

Assess business value vs. project complexity to prioritize candidate opportunities for generative AI

Prioritize opportunities/initiatives with high business value and low project complexity

Prioritization criteria exercise 1: Assessing the Create Content capability

Assessing the Create Content capability

This opportunity is removed because it does not pass the organization/business criteria

Prioritization criteria exercise 2: Assessing the Content Production capability

Assessing the Content Production capability

This opportunity is accepted because it passes the organization's business, responsible AI, and project criteria

3.2 Communicate policies for AI use

1-3 hours

1. Ensure policies for usage align with the organization's business criteria, responsible AI guiding principles, and ability to deliver the projects prioritized and beyond.
2. Understand the current benefits as well as limits and risk associated with any proposed generative AI-based solution.
3. Ensure you consider the following:
1. What data is being shared with the application?
2. Is the generative AI application deployed on the public cloud? Can anybody access the data provided to the application?
3. Avoid using very technical, legal, or fear-based communication for your policies.

Generative AI policy for the Create Content capability

Aligning policies to direct the uses assessed and implemented is essential

Example

Many of us have been involved in discussions regarding the use of ChatGPT in our marketing and sales initiatives. ChatGPT is a powerful tool that needs to be used in a responsible and ethical manner, and we also need to ensure the integrity and accuracy of its results. Here is our policy on the use of ChatGPT:

• You are free to use generative AI to assist your searches, but there are NO circumstances under which you are to reproduce generative AI output (text, image, audio, video, etc.) in your content.

If you have any questions regarding the use of ChatGPT, please feel free to reach out to our generative AI team and/or any member of our senior leadership team.

Generative AI policy for the Content Production capability

These policies should align to and reinforce your responsible AI principles

Example

Many of us have been involved in discussions regarding the use of ChatGPT in our deliverables. ChatGPT is a powerful tool that needs to be used in a responsible and ethical manner, and we also need to ensure the integrity and accuracy of its results. Here is our policy on the use of ChatGPT:

• If you use ChatGPT, you need to assess the accuracy of its response before including it in our content. Assessment includes verifying the information, seeing if bias exists, and judging its relevance.
• Employees must not:
• Provide any customer, citizen, or third-party content to any generative AI tool (public or private) without the express written permission of the CIO or the Chief Information Security Officer. Generative AI tools often use input data to train their model, therefore potentially exposing confidential data, violating contract terms and/or privacy legislation, and placing the organization at risk of litigation or causing damage to our organization.
• Engage in any activity that violates any applicable law, regulation, or industry standard.
• Use services for illegal, harmful, or offensive purposes.
• Create or share content that is deceptive, fraudulent, or misleading or that could damage the reputation of our organization.
• Use services to gain unauthorized access to computer systems, networks, or data.
• Attempt to interfere with, bypass controls of, or disrupt operations, security, or functionality of systems, networks, or data.

If you have any questions regarding the use of ChatGPT, please feel free to reach out to our generative AI team and/or any member of our senior leadership team.

Phase 4

Build the Roadmap

Phase 1
1. Establish Responsible AI Guiding Principles

Phase 2
1. Assess Current Level of AI Maturity

Phase 3
1. Prioritize Candidate Opportunities
2. Develop Policies

Phase 4
1. Build and Communicate the Roadmap

4.1.1 Create the implementation plan for each prioritized initiative

1-3 hours

1. Build the implementation plan for each accepted use case using the roadmap template.
2. Assess the firm's capabilities with respect to the dimensions of AI maturity and target the future-state capabilities you need to develop.
3. Prepare by assessing the risk of the proposed use cases.
4. Ensure initiatives align with organizational objectives.
5. Ensure all AI initiatives have a defined value expectation.
6. Do not ignore subject matter experts on either the business or IT side. You will need to consider both.

Download the AI Maturity Assessment and Roadmap Tool

Target-state options

Identify the future-state capabilities that need to be developed to deliver your use cases

1. Build an implementation plan for each use case to adopt.
2. Assess if the current state of the AI environment can be leveraged to deliver the selected generative AI use cases.
3. If the current AI environment is not sufficient, identify the future state required that will enable the delivery of the generative AI use cases. Identify gaps and build the roadmap to address the gaps.

4.1.2 Design metrics for success

1-2 hours

Establish metrics to measure to determine the success or failure of each POC.

1. Discuss which relevant currently tracked metrics are useful to continue tracking for the POC.
2. Discuss which metrics are irrelevant to the POC.
3. Discuss metrics to start tracking and how to track them with the generative AI vendor.
4. Compile a list of metrics relevant to the POC.
5. Decide what the outcome is if the metric is high or low, including decision steps and relevant actions.
6. Designate a generative AI application owner and a vendor liaison.

Prepare by building an implementation plan for each candidate use case (previous step).

Include key performance indicators (KPIs) and metrics that measure the application's contribution to strategic initiatives.

Consider assigning a vendor liaison to accelerate the implementation and adoption of the generative AI-based solution.

Generative AI POC metrics - examples

You need to measure the effectiveness of your initiatives. Here are some typical examples.

GitHub Copilot POC business value - example

Quantifying the benefits of GitHub Copilot to demonstrate measurable business value

POC Results

Task 1: Creating a web server in JavaScript

• Time to complete task with GitHub Copilot: 1 hour 11 minutes
• Time to complete the task without GitHub Copilot: 2 hours 41 minutes
• Productivity Gain = (1 hour 30 minutes time saved) / (2 hours 41 minutes) = 55%
• Benefit per Programmer = 55% x (average salary of a programmer)
• Total Benefit of GitHub Copilot for Task 1 = (benefit per programmer) x (# of programmers)

Enterprise Value of GitHub Copilot = Total Benefit of GitHub Copilot for Task 1 + Total Benefit of GitHub Copilot for Task 2 + ... + Total Benefit of GitHub Copilot for Task n

Source: GitHub

4.1.3 Build your generative AI initiative roadmap

1-3 hours

The roadmap should provide a compelling vision of how you will deliver the identified generative AI applications by prioritizing and simplifying the actions required to deliver these new initiatives.

1. Leverage tab 4, Initiative Planning, in the AI Maturity Assessment and Roadmap Tool to create and align your initiatives to the key value driver they are most relevant to:
1. Transfer the results of your value and complexity assessments to this tool to drive the prioritization.
2. Assign responsible owners to each initiative.
3. Identify which AI maturity capabilities each initiative will enhance. However, do not build or introduce new capabilities merely to advance the organization's AI maturity level.
2. Review the Gantt chart to ensure alignment and assess overlap.

Download the AI Maturity Assessment and Roadmap Tool

Build your generative AI roadmap to visualize your key project plans

Visual representations of data are more compelling than text alone.

Develop a high-level document that travels with the project from inception through to executive inquiry, project management, and finally execution.

A project needs to be discrete: able to be conceptualized and discussed as an independent item. Each project must have three characteristics:

• Specific outcome: An explicit change in the people, processes, or technology of the enterprise.
• Target end date: When the described outcome will be in effect.
• Owner: Who on the IT team is responsible for executing on the initiative.

Info-Tech Insight
Don't project your vision three to five years into the future. Deep dive on next year's big-ticket items instead.

4.1.4 Build a communication plan for your roadmap

1-3 hours

1. Identify your target audience and what they need to know.
2. Identify desired channels of communication and details for the target audience.
3. Describe communication required for each audience segment.
4. List frequency of communication for each audience segment.
5. Create an executive presentation leveraging The Era of Generative AI C-Suite Presentation and AI Maturity Assessment and Roadmap Tool.

Generative AI communication plan

Well-planned communications are essential to the success and adoption of your AI initiatives

To ensure that organization's roadmap is clearly communicated across the AI, data, technology, and business organizations, develop a rollout strategy, like this example.

Example

Deliver an executive presentation of the roadmap for the business stakeholders

After you complete the activities and exercises within this blueprint, the final step of the process is to present the deliverable to senior management and stakeholders.

Know Your Audience

• Business stakeholders are interested in understanding the business outcomes that will result from their investment in generative AI.
• Your audience will want to understand the risks involved and how to mitigate those risks.
• Explain how the generative AI project was selected and the criteria used to help draft generative AI usage policies.

Recommendations

• Highlight the need for responsible AI to ensure that human-based requirements are being addressed.
• Ensure your generative AI team includes both business and technical staff.

Download The Era of Generative AI C-Suite Presentation

Bibliography

"A pro-innovation approach to AI regulation." UK Department for Science, Innovation and Technology, March 2023. Web.

"Artificial Intelligence Act." European Commission, 21 April 2021. Web.

"Artificial Intelligence and Data Act (AIDA)." Canadian Federal Government, June 2022. Web.

"Artificial Intelligence Index Report 2023." Stanford University, April 2023. Web.

"Automated Employment Decision Tools." New York City Department of Consumer and Worker Protection, Dec. 2021. Web.

"Bain & Company announces services alliance with OpenAI to help enterprise clients identify and realize the full potential and maximum value of AI." Bain & Company, 21 Feb. 2023. Web.

"Buzzfeed to use AI to write its articles after firing 180 employees." Al Mayadeen English, 27 Jan. 2023. Web.

"California Consumers Privacy Act." State of California Department of Justice. April 24, 2023. Web.

Campbell, Ian Carlos. "The Apple Card doesn't actually discriminate against women, investigators say." The Verge, 23 March 2021. Web.

Campbell, Patrick. "NIST Artificial Intelligence Risk Management Framework (AI RMF 1.0)." National Institute of Standards and Technology, Jan. 2023. Web.

"EU Ethics Guidelines For Trustworthy." European Commission, 8 April 2019. Web.

Farhi, Paul. "A news site used AI to write articles. It was a journalistic disaster." Washington Post, 17 Jan. 2023. Web.

Forsyth, Ollie. "Mapping the Generative AI landscape." Antler, 20 Dec. 2022. Web.

"General Data Protection Regulation (GDPR)" European Commission, 25 May 2018. Web.

"Generative AI Market: Global Industry Trends, Share, Size, Growth, Opportunity and Forecast 2023-2028." IMARC Group, 2022. Web.

Guynn, Jessica. "Bing's ChatGPT is in its feelings: 'You have not been a good user. I have been a good Bing.'" USA Today, 14 Feb. 2023. Web.

Hunt, Mia. "Canada launches data governance standardisation initiative." Global Government Forum, 24 Sept. 2020. Web.

Johnston Turner, Mary. "IDC's Worldwide Future of Digital Infrastructure 2022 Predictions." IDC, 27 Oct. 2021. Web.

Kalliamvakou, Eirini. "Research: quantifying GitHub Copilot's impact on developer productivity and happiness." GitHub, 7 Sept. 2022. Web.

Kerravala, Zeus. "NVIDIA Brings AI To Health Care While Protecting Patient Data." eWeek, 12 Dec. 2019. Web.

Knight, Will. "The Apple Card Didn't 'See' Gender-and That's the Problem." Wired, 19 Nov. 2019. Web.

"OECD, Recommendation of the Council on Artificial Intelligence." OECD, 2022. Web.

"The National AI Initiative Act" U.S. Federal Government, 1 Jan 2021. Web.

"Trustworthy AI (TAI) Playbook." U.S. Department of Health & Human Services, Sept 2021. Web.

Info-Tech Research Contributors/Advocates

Joel McLean
Executive Chairman

David Godfrey
CEO

Gord Harrison
Senior Vice President, Research & Advisory Services

William Russell
CIO

Jack Hakimian
SVP, Research

Barry Cousins
Distinguished Analyst and
Research Fellow

Larry Fretz
Vice President, Industry Research

Tom Zehren
CPO

Mark Roman
Managing Partner II

Christine West
Managing Partner

Steve Willis
Practice Lead

Yatish Sewgoolam
Associate Vice President, Research Agenda

Rob Redford
Practice Lead

Mike Tweedie
Practice Lead

Neal Rosenblatt
Principal Research Director

Jing Wu
Principal Research Director

Irina Sedenko
Research Director

Jeremy Roberts
Workshop Director

Brian Jackson
Research Director

Mark Maby
Research Director

Stacey Horricks
Director, Social Media

Sufyan Al-Hassan
Public Relations Manager

Sam Kanen
Marketing Specialist