The core mistake is treating AI as a technology risk. AI amplifies existing risks across many areas at once: employment, product liability, intellectual property, regulatory compliance, reputation and directors' liability. Boards also go wrong by treating AI as an IT project, accepting management reassurance without evidence, confusing AI spending with governance, ignoring vendor AI, and waiting for regulation to force action. Better oversight means a named committee, regular risk reporting, a full AI inventory and one accountable executive.
Free AI leadership role brief
Before you write a job description, get a one-page brief on your role after a 30-minute call: scope, reporting line, a pay range from public filings, and where the candidates are.
Get my role brief →Why is "AI is a technology risk" a category error?
AI doesn't create one new risk to file next to cybersecurity. It overlays and amplifies existing ones. A biased hiring outcome is at once an employment-law risk, a compliance risk, a reputational risk and potentially a directors' liability risk. A generative AI tool that invents facts in a customer-facing product touches product liability, professional liability and brand risk. Boards that leave AI to the technology committee alone are organizing oversight around the wrong category. The risk is organizational, not just technical.
Why are expectations rising?
Regulation is becoming concrete: the EU AI Act's high-risk obligations apply from December 2, 2027; US states keep adding AI rules; and US banking agencies replaced their model risk guidance with SR 26-2 in April 2026. Directors are increasingly expected to know where AI is used, to make sure governance exists and to show that material risks were identified and addressed. The standard is moving faster than most boards.
What are the five common mistakes?
1. Treating AI as an IT deliverable
AI governance needs legal, compliance, HR, operations and the board, not just the CIO. Run as a technology workstream, its legal and ethical dimensions get neglected.
2. Relying on management reassurance
Be wary of presentations that cover AI opportunity but not risk. "We have appropriate controls" without documentation, measures or independent validation is trust without verification.
3. Confusing AI investment with AI governance
Spending on platforms, tools and talent isn't governance. Organizations focused on adoption speed often have the weakest deployment discipline.
4. Ignoring third-party AI
Most organizations use far more vendor AI than they build. Every SaaS product with AI features adds risk. Ask what due diligence is done on AI vendors and whether contracts cover model updates, data handling and incident response; under the EU AI Act, deployers keep their own duties whatever the vendor promises.
5. Waiting for regulation
Waiting for one definitive rule guarantees you're behind when it arrives. The obligations are already arriving piece by piece, and litigation over AI failures is a growing source of directors' exposure.
What does better oversight look like?
- A committee or working group with explicit responsibility for AI oversight.
- Regular reporting that includes risk indicators as well as opportunity measures.
- A complete AI inventory with risk classifications, requested by the board.
- A senior executive with real authority and resources who owns AI governance.
- AI treated as a standing discipline, not a one-off agenda item.
Five questions every board should ask about AI is a practical starting point, and what to have on file before the next audit lists the records.
What are other boards disclosing about AI oversight?
In proxy statements filed with the SEC from January to October 2026, 600 of 4,667 (12.9%) say the board or a board committee oversees AI. Of the 319 that name a committee, 189 (59%) give the job to the audit committee, and only 6 name a committee after AI. 457 (9.8%) describe a director with AI experience or expertise. See our full table of 2026 proxy disclosures.
Frequently asked questions
Is AI a technology risk?
Not mainly. AI amplifies existing risks across employment, product liability, intellectual property, regulatory compliance, reputation and directors' liability, so oversight can't sit with the technology committee alone.
What do boards get wrong about AI oversight?
Five common mistakes: treating AI as an IT project, accepting management reassurance without evidence, confusing AI spending with governance, ignoring vendor AI, and waiting for regulation before acting.
What does good board oversight of AI look like?
A committee or working group with explicit AI responsibility, regular risk reporting, a complete AI inventory with risk classifications, one accountable executive with real authority, and AI on the agenda as a standing item.
Written by Mariya with research and drafting help from AI tools. Every figure was checked against the sources below.
Sources
- Regulation (EU) 2024/1689 (EU AI Act), EUR-Lex.
- Federal Reserve, SR 26-2, Revised Guidance on Model Risk Management, April 17, 2026.
- EU AI Act high-risk dates as amended by the 2026 Digital Omnibus on AI: December 2, 2027 (Annex III) and August 2, 2028 (Annex I).
- SEC EDGAR DEF 14A filings, January to October 2026, compiled in Talent Echo, Board AI oversight in 2026 proxy statements.