In the LoopIssue 3, 1 May 2026All issues →

Boards must prepare for AI-related legal risk

Mobley v. Workday shows that AI liability does not stay with the supplier. Boards need one accountable owner for AI governance, and they need it now.

Most boards still assume that AI liability sits safely with the supplier. I think that'll be the most expensive mistake they make over the next year.

What Mobley v. Workday changed

The warning sign is Mobley v. Workday. The case started as an age discrimination claim against Workday's AI hiring algorithm, but worryingly, it didn't stop with Workday. Dozens of companies that used the tool were also pulled in as co-defendants. The court was prepared to see Workday as an agent of those companies, not a separate supplier sitting safely outside the legal blast radius.

This cuts through one of the most comfortable assumptions in enterprise AI. Many businesses have been signing vendor contracts as if they are buying insulation. They're not. If an AI tool rejects a candidate, mishandles data, creates a compliance breach, or makes a decision that crosses a legal line, the company using it may still have to answer for it.

Vendor contracts may help manage AI risk, but they won't remove responsibility when automated decisions cause legal or commercial harm.
Vendor contracts may help manage AI risk, but they won't remove responsibility when automated decisions cause legal or commercial harm.

One accountable owner for AI governance

This should prompt companies to take AI governance seriously. Not as a policy document. Not as a committee. Not as something owned vaguely by legal, compliance, IT and procurement at the same time. Businesses need one accountable owner for AI governance who can be asked to their face what went wrong.

That sounds blunt, but the alternative is worse. When everyone owns AI risk, no one does. A working group can review a framework. It can't take responsibility when a tool bought nine months ago starts creating regulatory exposure across hiring, sales, customer support, finance or HR.

Regulation is outpacing enterprise readiness

The pace of regulation is now moving too quickly for that old model. AI mentions in global legislative proceedings have grown ninefold since 2016 and risen another 21.3% in 2024 alone. In the US, 145 AI laws were enacted last year. The legal system is waking up to AI faster than most enterprises are waking up to their own exposure.

And the preparedness gap is concerning. Only 3% of compliance professionals say their organisation is ready for AI regulation as things stand. That means companies are deploying tools into live business processes while still unclear on who owns the risk and what happens when the output causes harm.

When everyone owns AI risk, no one does.
John Margerison

I see this all the time when talking to enterprise teams. The first question is usually about what AI can do. The second is about speed. The question that comes too late is the one that matters most: what could go wrong, and who is accountable when it does?

Audit the AI tools you already use

None of the fixes are difficult. They're just not particularly popular. Businesses need to audit every existing AI tool, not only the new ones they plan to buy. They need to look again at the contracts they rushed through during the first wave of adoption. And they need to ask the questions that never came up at signing.

Where is the data processed? Who reviews the output? Can the system explain why it made a decision? What happens if it produces biased, unlawful or commercially damaging results? Who has the power to stop it?

That isn't anti-AI. It's how AI earns trust. Real adoption doesn't come from giving agents more freedom and hoping the contract protects you. It comes from controlled automation, visible audit trails, clear ownership and humans approving the final step in high-stakes decisions.

Mobley v. Workday will make boards uncomfortable. I think that's a good thing. It's better to be uncomfortable now than surprised in court later.

By 2027, I expect many enterprises will face AI lawsuits they thought were someone else's problem. To avoid the worst of it, companies need to make sure they can prove who owns AI governance, how their systems are controlled, and why their people stayed in the loop before damage was done.

Keep reading. It's free

Still to come in this issue

  • One accountable owner for AI governance
  • Regulation is outpacing enterprise readiness
  • Audit the AI tools you already use

2 more minutes to read.

Read by 1,000+ founders & CEOs worldwide. No spam, ever. One click to unsubscribe.

The reading list

Three things John has been reading, and why they're worth your time.

  1. businessquarter.co.ukAI liability warning reaches enterprise boards

    Business Quarter has covered my thoughts on what the recent Mobley v. Workday case means for enterprises that could soon be facing a wave of AI-related legal action.

  2. ReutersEU lawmakers clinch provisional deal on watered-down AI rules

    EU countries and lawmakers reached a provisional deal to water down parts of the bloc's AI rules, including delaying high-risk system obligations until late 2027.

  3. n.comWipro share price spikes nearly 5% after IT firm's AI partnership with ServiceNow

    Wipro shares rose after the company expanded its partnership with ServiceNow to integrate agentic AI workflows into IT, HR, procurement and cybersecurity.

Questions this issue answers

Why does Mobley v. Workday matter for companies that use AI hiring tools?

AI management expert John Margerison explains that the case began as an age discrimination claim against Workday's AI hiring algorithm, but dozens of companies using the tool were pulled in as co-defendants. The court was prepared to treat Workday as an agent of those companies rather than a separate supplier. He argues this shows the business using an AI tool may still have to answer for its decisions.

Who should be responsible for AI governance in a business?

AI management expert John Margerison argues that businesses need one accountable owner for AI governance who can be asked directly what went wrong. He warns against AI risk being owned vaguely by legal, compliance, IT and procurement at once, because when everyone owns AI risk, no one does. A working group can review a framework but cannot take responsibility when a tool creates regulatory exposure.

What should boards do now to reduce AI-related legal risk?

AI management expert John Margerison says businesses should audit every existing AI tool, revisit contracts rushed through during early adoption and ask questions that never came up at signing. These include where data is processed, who reviews the output, whether the system can explain its decisions and who can stop it. He calls for controlled automation, visible audit trails, clear ownership and humans approving the final step in high-stakes decisions.

Written by

John Margerison

Founder and CEO, YesBusiness AI

Roughly monthly. Unsubscribe any time.

Get in touch

For press, speaking or working with John.