eBook

Chart the Course

for Responsible
AI Governance
Harnessing cloud data management to navigate
AI compliance in the wake of new regulations
Chart the Course for Responsible AI Governance

Contents
Foreword3 Part Three

Principles of Responsible AI 16

Introduction4
- Redefining Business with Generative AI 4 Data Leader's Checklist 18
- AI Data Management Needs 5
Part Four
Data Leader's Checklist  5
Informatica Intelligent Data Management Cloud 19
Part One - How IDMC Supports Responsible AI Governance  20

The Landscape of AI Legislation 7 Informatica’s AI Development Principles 22

- The EU AI Act 7
Data Leader’s Checklist 24
- Beyond EU Boundaries 9
- Singapore Model AI Governance Framework 9 Success Stories
- U.S. Executive Order on AI 9
- Other Regulations in Focus 9 Driving Better Business Outcomes Through Responsible
AI Data Governance 25
Data Leader’s Checklist 11
Conclusion27
Part Two

Challenges in AI Governance 12 Further Resources 28

- Data Integration 13
- Data Quality and Observability 13
- Data Bias 14
- Data Privacy 14

Data Leader’s Checklist 15

informatica.com 2
Chart the Course for Responsible AI Governance

Foreword
AI technology has rapidly moved from the edge of the adoption curve
to mainstream interest. This pace of change can be unsettling for
many enterprises, especially as they grapple with the reality of using
AI effectively and responsibly.

Numerous aspects are pivotal to this technological shift, with effective

data management at its core.

This eBook endeavors to help you chart a course for responsible AI

governance and what this means towards supporting compliance with
evolving AI regulations, understanding the challenges of this changing
legal landscape, and offering guidance for enabling integrated cloud
data management solutions to achieve your goals.

As part of this analysis, a checklist of actions outlined for each section

will support your enterprise in developing a data governance strategy
for responsible AI outcomes.

informatica.com 3
Chart the Course for Responsible AI Governance

Introduction
Redefining Business With Generative AI problems of bottlenecked data pipelines, data quality issues and
insufficient data management strategies. They will use GenAI as a
˝Generative AI (GenAI) is fundamentally changing the way we interact forcing function, taking long-needed actions to streamline, observe and
with technology — and the world. It is the fulcrum that businesses rely control their processes for managing multi-structured data.3
on to enhance, empower, and engage employees and customers. ̏¹

2024 has seen this AI evolution rapidly expand across many enterprises.
In the latest McKinsey Global Survey on AI, 65% of respondents report Organizations will need to demonstrate how they are
that their organizations regularly use GenAI. Looking ahead, most using AI safely and responsibly and be able to easily
respondents — 67 percent — expect their organizations to invest more in
answer why their AI models have made certain decisions.
AI over the next three years.2

A robust AI strategy is now essential for business. As companies move

up this adoption curve, they are met with a perennial challenge: data The onus is on organizations themselves to understand the heritage and
that is incomplete, inaccurate, insecure and therefore untrustworthy. lineage of their data inputs, ensure that the data utilized is high quality
This brings data management to the fore, as you can’t have an effective, and illustrate how AI performs to generate a particular outcome.
compliant AI strategy without reliable data.
“Only by taking this approach can they balance AI innovation with
GenAI presents companies with the appealing opportunity to generate compliance,” remarks EMEA Chief Architect Siddharth Rajagopal at
new business value and compete in dynamic markets more effectively Informatica. He also emphasizes the challenges organizations face
through increased business and operational automation. Innovative when integrating advanced AI technology within the frameworks of
companies will realize this opportunity by first tackling the persistent modern enterprise compliance.

¹ Generative AI Trends For All Facets of Business (forrester.com)

² https://www.mckinsey.com/featured-insights/mckinsey-global-surveys
³ How to create generative AI success with governed data management | Informatica

informatica.com 4
Chart the Course for Responsible AI Governance

AI Data Management Needs Data Leader’s Checklist

The pressure on enterprises globally to use AI for business advantage is
accelerating rapidly and that means understanding how this impacts the
Perform an external gap analysis to understand what data
business in real time to help ensure it adheres to changing regulations.
and AI governance solutions exist.
The size, complexity and distributed nature of data, combined with
increasing requirements for speed of action, mean the manual data
governance and management practices of the past can never keep up
Perform an external gap analysis to identify what methods of
with the agile business and legal needs of today.
manual governance are performed today that are unreliable
The underlying technology foundation supporting the execution of the and don’t scale.
data management strategy requires AI-powered automation to scale. An
integrated, cloud-native, modular platform can deliver support to de-risk
AI initiatives and enhance compliance with new AI regulations like the Identify missing key capabilities, such as robust security and
EU AI Act. privacy controls for handling sensitive data.

Establish reliable standards for data quality and ensure

To meet the demands of modern data and AI governance,
sufficient transparency.
a data and AI model management solution should
scale as the business grows and intelligently automate
governance and data management as a flexible,
Verify dependable policy enforcement that aligns with the
extensible, agile platform. data strategy for proper data and AI usage.

It is clear an innovative, robust solution that can deal with the

complexities of managing AI responsibly is essential.

informatica.com 5
Chart the Course for Responsible AI Governance

“Generative AI has the power to

be as impactful as some of the
most transformative technologies
of our time. The mass adoption
of generative AI has transformed
customer and employee
interactions and expectations. As
a result, GenAI has catapulted AI
initiatives from ‘nice-to-haves’ to the
basis for competitive roadmaps.”

Srividya Sridharan
VP and Group Research Director at Forrester

Source: Generative AI Trends For All Facets of Business (forrester.com)

informatica.com 6
Chart the Course for Responsible AI Governance

Part One

The Landscape of AI Legislation

As the adoption of generative AI systems explodes across enterprises European Union (EU).4 Its aim is to ensure the safety and fundamental
globally, then regulation must logically follow. rights of citizens and businesses. This Act classifies AI systems based
on their associated risks and implements specific rules for each
category, fostering the use of AI that is lawful, ethical and robust across
the region.
“Generative AI adoption is quickly moving forward,
and organizations are figuring out how to comply with The EU AI Act officially became law on August 1, 2024,5 setting a global
benchmark for AI regulation. The Act is intended to ensure that AI
regulation. Executives need to proceed with vigilance as
systems are:
regulation around generative AI is built and implemented.”
y Safe, transparent and traceable
Emily Frolick
y Non-discriminatory
Partner US Trusted Imperative Leader KPMG
y Environmentally friendly
y Respectful of existing privacy laws and people’s fundamental rights
Many countries are examining how best to manage AI regulation and the The Act establishes rules for creating and using AI in the EU, aiming to
first region to publish new legislation is Europe. foster innovation and curb potentially harmful uses. It states, “The AI Act
aims to provide AI developers and deployers with clear requirements
The EU AI Act and obligations regarding specific uses of AI. At the same time, the
The European Union Artificial Intelligence Act, also known as the EU AI regulation seeks to reduce administrative and financial burdens for
Act, is a comprehensive legal framework for AI regulation planned by the business, in particular, small and medium-sized enterprises (SMEs).”6

4 https://digital-strategy.ec.europa.eu/en/policies/regulatory-framework-ai
5 https://artificialintelligenceact.eu/documents/#:~:text=The%20AI%20Act%20will%20likely,between%20late%202024%20–%20summer%202027.
6 https://digital-strategy.ec.europa.eu/en/policies/regulatory-framework-ai

informatica.com 7
Chart the Course for Responsible AI Governance

The Act categorizes AI systems based on their risk to rights and safety.7
The Act defines the following:

1. Unacceptable risks, such as social scoring by governments,

will be banned.
2. High-risk systems, including those used in critical infrastructures,
law enforcement or procedures affecting workers' rights, must
comply with strict requirements before deployment. This includes
rigorous testing, transparency obligations and the provision of
high-quality data sets that avoid discrimination. A prevalent
example is real-time remote biometric identification systems,
like facial recognition.
3. AI systems deemed low or minimal risk, such as the use of chatbots,
will have lighter regulatory burdens; for example, AI-generated text
will have to be labeled as such and AI-enabled video games or spam
filters will have free use, encouraging innovation, while ensuring that
systems are trustworthy.

7 https://www.artificial-intelligence-act.com/#:~:text='Artificial%20intelligence%20system'%20
(AI,logic%2D%20and%20knowledge%20based%20approaches%2C

informatica.com 8
Chart the Course for Responsible AI Governance

Beyond EU Boundaries The proposed framework aims to create a trusted environment for
Understanding how to comply with the organizational requirements AI development globally, emphasizing principles like explainability,
of the EU AI Act naturally extends to operating in the market with both transparency and fairness.9
customers and competitors. As such, the impact of this new legislation Singapore’s AI governance framework emphasizes responsible and
will be felt beyond EU boundaries due to the global nature of the digital ethical AI development. The key principles are:
economy. It states, “The aim of the new rules is to foster trustworthy AI
in Europe and beyond by ensuring that AI systems respect fundamental 1. Transparency, Explainability and Fairness: Decisions made by AI
rights, safety and ethical principles and by addressing risks of very systems should be transparent, explainable and fair. This ensures that
powerful and impactful AI models.”8 users understand how AI arrives at its conclusions.
Therefore, the EU AI Act can compel non-EU countries and companies 2. Human-Centric Approach: AI solutions should prioritize human well-
to evaluate and potentially adjust their AI strategies and services if they being and safety. They should amplify human capabilities rather than
wish to operate in the EU market. This approach could standardize replace them.10
controls across regions, enabling some organizations to gain a
“Good governance is crucial. With the right guard rails in place, we create
competitive edge by aligning better with regulations and managing
conditions to innovate safely, responsibly and for a common purpose.
international partnerships more harmoniously. These far-reaching
The borderless nature of tech also means this must be a shared
implications reaffirm the importance of understanding and engaging with
endeavor,” said Deputy Prime Minister Heng Swee Keat. Singapore has
international regulations in an increasingly complex world. (You can read
recognized the global nature of AI innovation in enterprise usage and
more in the article here).
the need to work together across countries and regions to manage AI
regulatory compliance.
Singapore Model AI Governance Framework
The framework identifies nine areas — including accountability, trusted
Model AI Governance Framework for Generative AI: Singapore has
data for AI training and content provenance — where the governance of
developed a proposed framework for generative AI, which expands on
generative AI can be strengthened.
the existing Model Governance Framework that covers traditional AI.

8 https://digital-strategy.ec.europa.eu/en/policies/regulatory-framework-ai
9 https://www.imda.gov.sg/resources/press-releases-factsheets-and-speeches/press-releases/2024/public-consult-model-ai-governance-framework-genai
¹0 https://www.pdpc.gov.sg/help-and-resources/2020/01/model-ai-governance-framework

informatica.com 9
Chart the Course for Responsible AI Governance

U.S. Executive Order on AI There are seven AI-related bills in the California legislature,
Biden’s “Executive Order on the Safe, Secure and Trustworthy demonstrating the growing patchwork of U.S. state AI laws.
Development and Use of Artificial Intelligence”11 was published on Oct Research shows that efforts to regulate AI appear to be growing rapidly.
30, 2023. It provided U.S. government-authored guidance on eight Stanford University's 2023 AI Index shows that 37 AI-related bills were
primary sections: passed into law globally in 2022, with countries such as Canada, India,
y New Standards for AI Safety and Security China and Switzerland all developing AI regulations.12

y Protecting Americans’ Privacy As illustrated here, many countries are in various stages of evolution in
regulating AI and have differing views of how best to do it. Consequently,
y Advancing Equity and Civil Rights
there is no clear consensus on how best to regulate AI either from
y Standing Up for Consumers, Patients and Students a global standpoint or consistently, which allows innovation while
y Supporting Workers curtailing potentially harmful uses of AI.

y Promoting Innovation and Competition As EMEA Informatica Chief Architect, Siddharth Rajagopal analyzes the
global landscape, and he predicts, “We will likely see a divergence in
y Advancing American Leadership Abroad
how countries across the globe regulate AI. This could look like different
y Ensuring Responsible and Effective Government Use of AI variants of the same regulations that are loosely based on the EU AI Act.
The order is an initial step to ensure that AI is trustworthy and helpful. But large, multi-national organizations will not be able to afford to do
It aims to guide how AI is developed so that companies can profit AI regulation on a siloed project-by-project, country-by-country basis. It
without jeopardizing public safety. (You can read more about it in is too complex.” Rajagopal further emphasizes the need for a strategic
an article here.) approach, stating, “Instead, they will need to consider how AI regulation
translates into policy and put solid foundations in place that can be
Other Regulations in Focus easily adapted for individual regions. For example, regulators across
countries are showing an appetite for transparency.”
In May 2024, the Colorado Artificial Intelligence Act (CAIA) became law
in the U.S., making Colorado the first state to enact legislation governing
the use of high-risk artificial intelligence systems.

¹¹ https://www.whitehouse.gov/briefing-room/presidential-actions/2023/10/30/executive-order-on-the-safe-secure-and-trustworthy-development-and-use-of-artificial-intelligence/
¹² https://hai.stanford.edu/news/2023-state-ai-14-charts

informatica.com 10
Chart the Course for Responsible AI Governance

Data Leader’s Checklist

Review and understand scope, implications and requirements for navigating the latest
AI regulations, such as the EU AI Act, with key stakeholders across the enterprise.
Based on their understanding, senior leadership must define clear objectives for
compliance that align with both the legal requirements and the company’s strategic
goals. This could include adjusting current AI systems, defining new AI data
governance structures or implementing specific data protection measures.

Develop a detailed risk management framework that addresses the specific exposure
associated with unreliable AI and data management practices or unknowns (e.g., user
access rights).

Establish and communicate clear data and AI policies regarding data usage, AI model
deployment and regulatory compliance.

informatica.com 11
Chart the Course for Responsible AI Governance

Part Two

Challenges in AI Governance
Many companies today have inefficient data management
processes that result from disorganized pipelines, inadequate data
observability and a lack of governance controls. These perennial
problems can make generative AI outputs unfit for use, which could
lead to poor decisions, frustrated customers and significant policy
compliance violations.

The growing number of countries and regions enforcing new AI

regulations demonstrates that CDOs must understand how to comply
with an array of new laws across relevant territories where they do
business. This could be a bewildering checklist of risk management
challenges for international organizations operating across borders. As
such, the evolving nature of generative AI means that businesses must
stay abreast of relevant laws and regulations, including AI compliance
laws as well as intellectual property rights and data protection, to
ensure compliance and mitigate legal risks. Robust data governance
that ensures the availability and security of high-quality, reliable data
throughout the enterprise is essential.

This complex scenario illustrates some of the crucial pillars of modern

data governance, including managing data and AI risk and compliance,
data quality and observability, data sharing and democratization, and
cataloging and lineage for integration and transparency.

Let’s take a closer look at these capabilities and how they support
modern data governance.

informatica.com 12
Chart the Course for Responsible AI Governance

Data Integration Data Quality and Observability

Navigating responsible data and AI governance requires a modern “Data quality has always been an important issue for CDOs,” observed
approach supported by a cloud data management platform. McKinsey in The data dividend: Fueling generative AI, 2023. “But the
scale and scope of data that generative AI models rely on has made
the ‘garbage in/garbage out’ truism much more consequential and
expensive, as training a single LLM can cost millions of dollars.”14
Among data leaders implementing or planning to
implement generative AI, a recent Informatica CDO There are two core challenges to ensuring high data quality: extending
data observability programs for generative AI applications to better
Insights 2024 Survey reveals, the quality of data (42%)
spot quality issues, such as by setting minimum thresholds for
is the top data-related obstacle to the adoption of
unstructured content to be included in generative AI applications; and
generative AI and large language models (LLMs). And
developing interventions across the data life cycle to fix issues.
data sources are expanding as well: 41% admit they’re
juggling 1,000 or more data sources — a number 79% Moreover, AI can produce inconsistent, inaccurate and even biased
outputs without a robust data governance framework to confidently
expect to increase during 2024.13
deliver data to:

y Maintain data quality and provide clean, consistent and accurate

datasets, and in turn, reduce noise, improve model accuracy and
The disparate nature of data sources in use in modern technology
support sound decision-making.
networks means that data could be spread across silos due to legacy,
fragmented data governance technologies that require significant y Safeguard data exposure by aligning data usage with data
manual intervention and hinder the company’s ability to fuel AI protection standards and managing data access as appropriate to
initiatives with trusted data. Moreover, embracing AI requires shifting the AI use case.
from legacy on-premises data governance systems towards an agile, y Provide transparency and explainability into how AI models use
scalable and cost-effective cloud-native solution to keep up with data to build trust among users and hold entities accountable
growing data demand. for AI outcomes.

¹³ CDO Insights 2024: Charting a Course to AI Readiness

¹4 https://www.mckinsey.com/capabilities/mckinsey-digital/our-insights/the-data-dividend-fueling-generative-ai

informatica.com 13
Chart the Course for Responsible AI Governance

Data Bias This leads to ethical considerations of how to manage data responsibly,
ensuring data is handled according to defined ethical factors such
as transparency, fairness and bias when applying data management
policies for appropriate use.
“The danger with AI is that it is designed to work at
scale, making huge numbers of predictions based on Data Privacy
vast amounts of data. Because of this, the effect of just
Data leaders surveyed in our CDO Insights 2024 survey are seeing
a small amount of bias present in either the data or the
roadblocks in AI governance. A notable 35% revealed ‘The inability
algorithms can quickly be magnified exponentially.” 15 to minimize risks related to industry regulations and compliance laws,17
Bernard Marr which could be complicating their frustration with their inability
Futurist Author, Keynote Speaker and Strategic Business Advisor to manage access to private and sensitive data (33%).18 Moreover,
45% saw data privacy and protection as a top data management
investment priority.

32% of CDOs in our Informatica “CDO Insights 2024: Charting a Course The lack of oversight in data privacy and protection can pose
to AI Readiness” survey saw bias as a top challenge in adopting or a more significant challenge and lead to a loss of trust in AI
planning to adopt GenAI technology.16 adoption, for example, when using personally identifiable information
(PII) inappropriately.
Bias can come in different forms, such as:
Enabling security controls while allowing data access for various
1. Statistical bias, which includes data and algorithmic bias. qualified teams and roles is a critical challenge in AI and data
2. Societal bias, which encompasses human and business bias that management, especially for maintaining compliance with security
policies. It's important to clarify data ownership and establish rights for
could be codified into business policies. Such biases can have severe
proper data use. Data should be managed with a clear understanding of
implications. Training AI systems on incomplete or biased data can
who owns or produces it, and how it should be used appropriately. It's
lead to incorrect patterns and, in turn, make incorrect predictions
crucial to provide data only to those who need it and are aligned to
or develop hallucinations — where AI systems generate false or
the data strategy, while respecting data ownership and use policies.
distorted outputs — resulting in incorrect or misleading results.

¹5 https://bernardmarr.com/building-responsible-ai-how-to-combat-bias-and-promote-equity/
¹6 CDO Insights 2024: Charting a Course to AI Readiness
¹7 Ibid.
¹8 Ibid.

informatica.com 14
Chart the Course for Responsible AI Governance

AI offers tremendous promise for data management potential, but aligning an entire organization
behind it will become increasingly difficult as data leaders contend with multiple priorities as well
as multiple roadblocks.

The need for a single consolidated solution to manage all this data becomes abundantly clear.
Relying on existing staff to upskill or reskill to manage AI, as 48% of data leaders from the
CDO Insights 2024 survey19 would consider, can allow them to make strides in achieving their
AI and data goals — but that won’t be enough without providing these employees with the
proper technology.

Data Leader’s Checklist

Audit existing data management practices by regularly reviewing and refining data
management processes to eliminate inefficiencies and disorganized data pipelines.

Upskill employees and provide data literacy programs. Consider rolling out continuous
education focusing on AI and data literacy for relevant staff. (You can read more about
how Informatica supports data literacy programs here).

Develop training programs on bias mitigation and data ethics for employees.

¹9 CDO Insights 2024: Charting a Course to AI Readiness

informatica.com 15
Chart the Course for Responsible AI Governance

Part Three

Principles of Responsible AI
Ethical frameworks are crucial to understand in order to guide AI development and deployment. Principles of responsible AI, listed below, provide
some guidelines to explain what it means from our perspective. This foundation is vital for building sustainable AI practices that are compliant and
meet societal expectations.

Here is a brief overview of each one:

Fairness and Bias Mitigation Transparency and Explainability

Fairness: AI systems should provide equitable outcomes for all users, Transparency involves the processes by which AI systems are
regardless of their background or demographics. It's about treating all designed and deployed, which includes clear disclosure about AI's
groups fairly and embedding ethical considerations in AI systems. decision processes and capabilities. Organizations must discover and
catalog their most relevant data and metadata to certify the relevance,
Bias Mitigation: Efforts must be made to detect and reduce or eliminate
value and security and to ensure transparency.
biases in AI systems that can lead to unfair treatment of individuals
or groups. Have we removed personally identifiable information (PII) Explainability is the ability of AI systems to detail, in an understandable
and followed all regulations? Are we transparent, and can we prove the way, the processes and factors they use to make a decision.
lineage of the data the model uses? Can we document and be ready to Organizations must address the challenge of making AI techniques
show regulators or investigators that the data has no bias? more explainable and transparent. This helps build trust and supports
human oversight.

informatica.com 16
Chart the Course for Responsible AI Governance

Privacy and Data Protection Safety and Security

Privacy requires ensuring that personal information is kept secure Safety involves designing and implementing systems that reliably
and is only used in ways to which people have agreed. This means handle data without errors that could lead to harmful decisions. This
that organizations can detect data misuse, protect sensitive data and requires robust testing and validation to identify and mitigate potential
demonstrate that appropriate controls are maintained. risks associated with data handling, processing and decision-making.
Moreover, safety protocols ensure that AI systems are resilient against
Data Protection entails the legal and secure handling of data. It focuses
data corruption, unauthorized access and other vulnerabilities that could
on compliance to protect individuals’ data from unauthorized access
impair their functionality or compromise data integrity.
and breaches, and ensures policies are documented and enforced to
comply with privacy laws and regulations. Security entails protecting AI systems from malicious attacks and
ensures that they are reliable and resilient in the face of attempts to
Accountability and Governance tamper with or manipulate their operation. It also aligns security controls
with governance policies for enhanced protection.
Accountability means that the developers, operators and deployers of
AI systems are responsible for the consequences of their operation,
regardless of AI's autonomous capabilities. Being able to define and
Societal Impact
assign ownership of policies ensures accountability. When these policies Concerns include understanding and managing the broad effects of AI
can be broken into rules, then data quality measurement and reporting technology on society, such as ethical standards, societal benefit and
processes can be automated in the organization. avoiding potential negative consequences on social structures.

Governance establishes the frameworks, policies and standards to These principles help guide the creation and usage of AI technologies
manage AI development and deployment responsibly, focusing on that are beneficial for all stakeholders involved and are supportive of
ensuring compliance with ethical and legal standards. Data needs to general societal standards and values. They are at the center of building
be processed and made usable, trustworthy and democratized for use effective, robust data governance strategies and employing AI systems
across the organization, while adhering to governance policies. This that have been developed with the same goals (which we will discuss
should be flexible and adaptable to changing circumstances. further in the section on “Informatica’s AI Development Principles” for
product strategy).

informatica.com 17
Chart the Course for Responsible AI Governance

Data Leader’s Checklist

Oversee the implementation of fairness, transparency
and accountability measures.

Enforce privacy and data protection principles across

the organization.

Regularly audit and update governance frameworks to ensure compliance

and societal impact considerations. (You can read more about Informatica
governance frameworks here.)

Promote a culture of data-driven decision-making, continuous learning

and adaptation to AI-driven technologies.

informatica.com 18
Chart the Course for Responsible AI Governance

Part Four

Informatica Intelligent Data Management Cloud

An end-to-end solution for implementing essential data strategies and effective AI compliance

Informatica’s vision for data management is one of unification and empowerment. It encompasses a foundation enabling self-serve access to valid,
trustworthy data, with AI-infused tools amplifying productivity and enriching user experiences.20 Informatica Intelligent Data Management Cloud™
(IDMC) plays a central role in de-risking AI initiatives and enhancing compliance with regulations such as the EU AI Act. As shown in Figure 1, the
comprehensive solution offers data governance with privacy controls, data quality improvement and AI-powered data cataloging to ensure the
transparency, reliability and integrity of data. This is all done through a data management platform that is multi-vendor, multi-cloud (including AWS,
Azure and Google) and hybrid, supporting on-premises and cloud-based data storage and management. Through automated data management tasks
and seamless data integration, IDMC increases operational efficiency and creates a single source of truth.

Figure 1: Intelligent Data Management Cloud (IDMC) delivers what you need to achieve better business outcomes, all in one place.

²0 Designing a Principled Product Strategy for AI-Powered Data Management

informatica.com 19
Chart the Course for Responsible AI Governance

Leveraging CLAIRE®, Informatica’s unified metadata intelligence of 40 Increase Productivity Across the Organization
petabytes, introduces a natural language (NL)-based experience to
Improve quality and transparency of data by automating repetitive tasks
the IDMC. This dramatically simplifies tasks such as data discovery,
such as debugging, testing, refactoring and documentation. CLAIRE GPT
integration, quality, governance and master data management. The
helps data engineers quickly explore the datasets and find data quality
Informatica CLAIRE GPT, a generative AI-powered version of our
issues. It also helps document and test the pipeline.
pioneering AI engine CLAIRE, will enhance data management and data
utilization experience for all users. This rapidly improves a company’s Streamline Data Pipeline Creation With Fully Automated Workflows
time-to-value and amplifies business insights. Search and discover
CLAIRE GPT helps find the right datasets needed for the task directly
data assets are available in the data catalog through a natural language
from requirement specifications and generates the first draft of the data
interface powered by Informatica's large language model (LLM).
pipeline. CLAIRE GPT can create data mappings, data quality rules and
CLAIRE GPT helps simplify, accelerate and optimize data management
governance artifacts, thus allowing the data team to refine
operations, driving enormous gains in productivity for data teams.
them easily.

How IDMC Supports Responsible AI Governance These features enhance an already powerful Data Catalog as part
of IDMC. By harnessing CLAIRE's advanced machine learning and
Key benefits include:
AI capabilities, Data Catalog effectively ensures that data across
Empower Data Users and Reduce Data Management Costs the enterprise is consistently managed and compliant with relevant
regulations. Customers can automate the linkage of governance policies
CLAIRE GPT is a simple and easy-to-use interface. It allows more
directly to their enterprise data, ensuring that compliance (managing
business or citizen users to perform data management tasks without
relevant AI regulations) and best practices are embedded throughout
specialized knowledge, thus reducing the barrier to entry.
their data management processes (as illustrated in Figure 2 on
the next page).

informatica.com 20
Chart the Course for Responsible AI Governance

Figure 2: Example of an AI compliance dashboard from the Informatica Cloud Data Governance and Catalog, illustrating the management
of AI regulations.

This capability streamlines governance and enhances operational efficiency, allowing organizations to maintain a high standard of data integrity
and security without manual oversight. Organizations must have full trust and confidence in the applications they employ and be able to trace the
automation of data usage and governance from source to output.

informatica.com 21
Chart the Course for Responsible AI Governance

Informatica’s AI Development Principles

The development of the Intelligent Data Management Cloud incorporates AI, such as the CLAIRE engine, to automate processes and enable
predictive insights. Adhering to a stringent set of principles enhances platform reliability at scale. These principles are designed to ensure that what
is created and deployed using IDMC for AI is developed and used in a way that respects human rights, contributes to societal benefits, upholds
privacy and security and prioritizes transparency and understandability. They also strive for inclusivity and diversity.

These principles include:

y Enhancing human productivity in data management: We aim to y Providing transparency and understandability: We aim to create
develop AI technologies for data management, making it easy for AI models that are effective and understandable. We will leverage
data teams and business users to manage their data effectively. By advanced interpretability frameworks and tools to provide insights
narrowing our focus, we aspire to deliver impactful solutions while into how our models make decisions, allowing users to understand,
tailoring our technologies to the unique needs and challenges within where appropriate, how our AI application reached its conclusions.
this area. y Designing delightful user experiences: We aim to harness AI to
y Ensuring data security and accountability: We pledge to create augment human productivity by crafting thoughtfully designed user
AI technologies that prioritize data privacy and security and experiences that delight end users.
balance them against the functionality of our product features. AI y Democratizing AI responsibly: We commit to making AI accessible
development oversight will include third-party audits, robust feedback to a broad range of users while maintaining a strong focus on ethical
mechanisms and a dedicated oversight team. We will maintain and privacy considerations. We will balance openness with robust
documentary evidence of how our AI was trained. This will help to control mechanisms designed to prevent misuse of technology and
ensure transparency in our processes and trust in our operations. protect data privacy.

(You can read more about Informatica’s vision in this blog article, “Designing a Principled Product Strategy for AI-Powered Data Management.”)

informatica.com 22
Chart the Course for Responsible AI Governance

“Data cannot be an afterthought in generative AI.

Rather, it is the core fuel that powers the ability of
a business to capture value from generative AI.” 21

McKinsey, The data dividend: Fueling generative AI

21 https://www.mckinsey.com/~/media/mckinsey/business%20functions/mckinsey%20digital/our%20insights/the%20data%20dividend%20fueling%20generative%20ai/
the-data-dividend-fueling-generative-ai.pdf?shouldIndex=false

informatica.com 23
Chart the Course for Responsible AI Governance

Data Leader’s Checklist

Conduct a high-level assessment of current data governance capabilities and identify
gaps that can be addressed by an advanced solution like IDMC.

Develop a detailed strategy that incorporates AI capabilities to manage, discover

and govern data. This strategy should align with overall business objectives and
technological landscapes.

Prepare a change management plan to address potential human and

process modifications.

informatica.com 24
Chart the Course for Responsible AI Governance

Success Stories

Driving Better Business Outcomes Through

Responsible AI Governance
Informatica has worked with customers data models, leading to significant data management capabilities have allowed
worldwide to help them make better compliance challenges. Limited visibility the company to transition to a modern data
business decisions through responsible into the origins and consumption of data mesh architecture. This architecture features
AI-powered data governance. These two increases the risk of compliance oversight near real-time data availability for analysis
examples from the healthcare and telecom and undermines the reliability of business and advanced analytics, achieved through
industries demonstrate our longstanding reporting. Non-compliance with data privacy standardized, governed and automated
commitment to strategic partnerships that regulations, such as GDPR and HIPAA, could data flows.
deliver through technological advances. result in substantial fines and damage
to the company's reputation. Additionally, Outcomes:
Biopharma’s Swift AI fragmented tools and varying technologies The organization has significantly accelerated
Compliance Upgrade across departments have led to inconsistent its generation of insights through the easier
A multinational pharmaceutical and data management processes, further availability of trustworthy, safe and secure
healthcare company wanted to realize their elevating the exposure to compliance risks. data. Robust compliance measures have
vision of using AI to develop and market new enhanced the integrity and reliability of this
Solution: data, improving the accuracy of AI-generated
drugs more quickly, optimize their product
portfolio and become the first biopharma By leveraging the capabilities of Informatica insights. Additionally, AI-driven automation
company to implement responsible AI on a Intelligent Data Management Cloud, has bolstered the organization’s capacity to
large scale. Biopharma now has a unified and accurate govern disparate data effectively and adapt
view of data across the organization, enabling quickly to changing AI regulations, all while
Challenges: a connected data landscape that supports the reducing technology costs.
The organization faced a complex data and delivery of trusted, high-quality and compliant
AI environment with over 150 disparate data to AI applications. Integrated AI-powered
Chart the Course for Responsible AI Governance

AI-Driven Data Unification in Telecom cross-functional use cases. Additionally, products, making them globally accessible
multiple governance standards across to partners and customers. Additionally,
A leading European multinational
different geographies and business entities Informatica has facilitated timely access to
telecommunication company faced a growing
contribute to operational inefficiencies. high-quality,
need to integrate diverse data sources
The company also faced continuous and cross-functional data, significantly
from various regions to responsibly cater to
escalating contractual and regulatory accelerating the transformation of the
different stakeholders and use cases. The
demands concerning the use of data and AI, end-to-end customer journey.
primary goal is to leverage data as a strategic
adding further complexity to its challenges.
asset across the enterprise to enhance
Outcomes:
business outcomes.
Solution: The implementation of Informatica's
Challenges: To address its data challenges, the solutions has led to readily available and
organization partnered with Informatica to trusted data, drastically reducing the time to
Navigating a vast and intricate data
implement a federated data governance realize business value. This enhancement
environment, characterized by over 1000
model. This model adheres to the FAIR in data accessibility and reliability has
enterprise applications, the organization had
principles, ensuring that data products also strengthened the organization’s
a critical need to govern diverse data
are findable, accessible, interoperable and adherence to contractual and regulatory
spanning enterprise networks and partners.
reusable. Informatica's solutions have compliance requirements, solidifying its
The company struggles with a scarcity of
empowered the company to democratize standing in a competitive and fast-evolving
reliable, reusable data assets that can support
high-quality, governed and secure data telecommunications industry.
Chart the Course for Responsible AI Governance

Conclusion
It is evident that using AI effectively requires a strategic approach 3. Appropriate Data Sharing – Enable responsible data and AI usage
to data governance coupled with compliance for evolving global with controls that align owners and producers to data and AI
regulations. According to Informatica’s CDO Insights 2024 Survey, consumption. This involves adopting data management practices
the top data strategy priorities for 2024 include delivering reliable that are transparent, fair and accountable, which are crucial for
and consistent data fit for generative AI (39%), improving data-driven maintaining public trust and meeting regulatory standards.
culture and data literacy (39%) and improving governance over data By leveraging the checklist in each section of this eBook,
and data processes (38%).22 These priorities all highlight the need to organizations can examine their readiness to adopt AI technologies
push beyond implementing and toward maximizing AI’s efficacy and across the business. These technologies support the delivery of
making the most of an organization’s investments. effective data governance using high-quality, compliant and
In summary, there are several key takeaways from this analysis: managed data.

1. Data and AI Governance for Regulatory Compliance – Have As AI and data management continue to advance, data leaders must
controls in place to map data strategy to mandates for appropriate recognize that they both go hand in hand, complementing each other
use policies, such as the new regulations like the EU AI Act. Staying to drive innovation. By harnessing their power collectively, leaders
ahead in regulatory compliance is crucial for enterprises to leverage can revolutionize their strategies and pave the way for transformative
new market opportunities without falling foul of regulatory non- progress within their organizations.
compliance, poor business decision-making, a tarnished reputation
and loss of customer trust.
2. Data Quality and Observability – Enhance the reliable performance
of data fueling GenAI and other applications such as analytics, by
providing transparency and explainability into how AI models use
data. This approach builds trust among users and holds entities
accountable for AI outcomes.

22 https://www.informatica.com/lp/cdo-insights-2024-charting-a-course-to-ai-readiness_4656.html

informatica.com 27
Chart the Course for Responsible AI Governance

Further Resources

Why AI-Powered Data Governance Helps Why AI-Powered Data Management Improves
Businesses Scale | Informatica GenAI Performance | Informatica

LEARN MORE LEARN MORE

informatica.com 28
Chart the Course for Responsible AI Governance

About Us
Informatica (NYSE: INFA), a leader in enterprise AI-powered cloud data management,
brings data and AI to life by empowering businesses to realize the transformative power
of their most critical assets. We have created a new category of software, the Informatica
Intelligent Data Management Cloud™ (IDMC), powered by AI and an end-to-end data
management platform that connects, manages and unifies data across virtually any
multi-cloud, hybrid system, democratizing data and enabling enterprises to modernize
their business strategies. Customers in approximately 100 countries and more than
80 of the Fortune 100 rely on Informatica to drive data-led digital transformation.
Informatica. Where data and AI come to life.™

Worldwide Headquarters
2100 Seaport Blvd.
Redwood City, CA 94063, USA
Phone: 650.385.5000
Fax: 650.385.5500
Toll-free in the US: 1.800.653.3871

informatica.com
linkedin.com/company/informatica
x.com/Informatica

CONTACT US
IN19-5005-0924

© Copyright Informatica LLC 2024. Informatica and the Informatica logo are trademarks or registered trademarks of Informatica LLC in the United
States and other countries. A current list of Informatica trademarks is available on the web at https://www.informatica.com/trademarks.html. Other
company and product names may be trade names or trademarks of their respective owners. The information in this documentation is subject to
change without notice and provided “AS IS” without warranty of any kind, express or implied.

informatica.com