AI is Making Life-Changing Choices—Can You Trust the Data Behind It?

When AI starts making life-altering decisions—who gets medical care, who qualifies for a mortgage, who gets a second chance—you can’t afford to let flawed, unchecked, or unethical data be the judge. 

Earlier this week, the International Association of Privacy Professionals released the Organizational Digital Governance Report, and a certain statistic grabbed my attention:

• 69% of chief privacy officers (CPOs) surveyed have been tasked with overseeing AI governance.
• The same 69% also shoulder responsibility for data governance and data ethics.
• But only 37% are tasked with cybersecurity regulatory compliance, and just 20% manage platform liability.

What this shows is a growing understanding that AI governance is inseparable from data governance. 

Working with companies like AAA Life Insurance, Norges Bank, and First American, we’ve learned firsthand that responsible data governance HAS to be at the heart of your AI strategy, particularly for insurance, finance, healthcare, and other highly regulated sectors where data accuracy and compliance are critical.

Beyond the Algorithms: Why Data Governance is the Real AI Foundation

When people think about AI governance, they usually focus on making sure AI models are transparent, accountable, and ethical. These are important conversations, no question. But none of these things can happen if your underlying data is a mess. This is especially true in industries like insurance, where AI models are determining underwriting risk, pricing policies, and assessing claims. The stakes are incredibly high, so the need for clean, accurate data to assess risk fairly is critical.

The same applies to financial services, where poor data governance could lead to biased credit decisions or incorrect loan approvals. Flawed data leads to flawed AI outputs—no matter how sophisticated the algorithms may be.

This is exactly where many companies stumble. They treat AI as a standalone concern, a new problem to tackle without realizing that AI governance is simply an extension of data governance. Data quality, security, privacy, and ethics—these aren’t separate issues for AI governance to worry about later. They are the foundation AI governance needs to work at all.

Data Mesh and the New Era of AI Governance

At Metaphor, we’ve long believed that data governance can no longer be an afterthought, especially as AI is here to stay. This is where our approach—centered around a collaborative data mesh architecture—comes in.

A data mesh emphasizes decentralized data ownership, meaning the teams that generate and use the data are responsible for governing it. This allows organizations to break down silos and eliminate bottlenecks, ensuring that AI models are using accurate, up-to-date data. 

For an insurance company, implementing a data mesh could result in improvements in the quality of data feeding their AI models, which allows for more accurate risk assessments and better policy pricing. This collaborative approach to data governance ensures that the AI systems are both transparent and trustworthy.

In other words, good AI governance flows naturally from good data governance—and the future of both is collaborative, decentralized, and adaptive.

The Overlap of Privacy, Ethics, and AI Governance

The other part of the IAPP report that stood out to me was how roles are evolving within organizations. We now see a clear overlap between privacy, data governance, and AI ethics responsibilities. This is no accident.

If AI models are making decisions about people’s lives—whether it’s determining who gets a loan, a medical diagnosis, or insurance coverage—you need to ensure those decisions are ethical, unbiased, and transparent. For example, without ethical data governance, a bank might see their AI-driven credit scoring models unintentionally reinforce biases. 

Without ethical data governance, these AI decisions become a black box—a recipe for disaster in terms of accountability and transparency.

Metaphor’s focus on data ethics by design—embedding ethical considerations into the very fabric of data governance—helps organizations move from reactive governance to proactive, responsible AI adoption, building AI systems that are trustworthy from the ground up.

Where Does Platform Liability Fit In?

The IAPP report noted that only 20% of CPOs are responsible for platform liability, which might sound surprising. But when you consider the connection between AI governance, data governance, and platform responsibility, you start to see the bigger picture.

If your AI models are making decisions within a platform—whether that's a financial app or an e-commerce marketplace—how those models behave is directly tied to liability. Faulty AI decisions trace back to bad data, which ultimately falls on the governance framework you’ve built (or failed to build). 

So, while platform liability may not yet fall squarely under CPO responsibilities, it’s clear that without robust AI and data governance in place, liability issues are just a matter of time.

The Metaphor Difference: Data Governance That Powers AI

At Metaphor, we see data governance as the backbone of AI innovation, especially in regulated industries like insurance, finance, and healthcare. We’ve worked with companies across sectors to ensure that their AI models are built on a foundation of reliable, well-governed data. From AAA Life Insurance to Norges Bank, our clients trust Metaphor to help them manage the massive volumes of data driving their AI systems.

Our data mesh approach flips traditional governance on its head by decentralizing control, so your data becomes more accessible, trustworthy, and ready to power AI initiatives. The result? Faster AI deployment, fewer governance nightmares, and an AI ecosystem built on ethics and transparency.

If you’re serious about AI governance, it all starts with your data. Because AI governance doesn’t begin with the algorithms—it begins with how well you govern your data. Get that right, and everything else will follow.

‍

About Metaphor

The Metaphor Metadata Platform represents the next evolution of the Data Catalog - it combines best in class Technical Metadata (learnt from building DataHub at LinkedIn) with Behavioral and Social Metadata. It supercharges an organization’s ability to democratize data with state of the art capabilities for Data Governance, Data Literacy and Data Enablement, and provides an extremely intuitive user interface that turns even the most non-technical user into a fan of the catalog. See Metaphor in action today!