Why must AI governance policies address third-party AI systems?

Question

Options

• ALiability may still attach to deployers
• BThird parties are exempt from regulation
• CThird-party models are always open source
• DExternal systems cannot cause harm

• A
  88% (21)
• B
  8% (2)
• C
  4% (1)

Explanation

Deployers of AI systems retain legal and regulatory liability even when the underlying model or service is built and operated by a third party. If a company integrates an external AI into their product and it causes harm-discriminatory decisions, data breaches, faulty outputs-regulators and courts can still hold the deployer responsible. This is why governance policies must extend to vetting, auditing, and contractually constraining third-party AI systems, not just internally built ones.

Why the distractors fail:

• B is the opposite of reality; third-party AI providers are increasingly subject to regulation (e.g., the EU AI Act imposes obligations on providers and deployers alike).
• C is factually false; many third-party models (GPT-4, Gemini, etc.) are proprietary and closed source.
• D is demonstrably false; external systems can and do cause real-world harm, which is precisely the concern driving governance requirements.

Memory tip: Think of it like a restaurant sourcing food from a supplier-if a customer gets sick, the restaurant can't escape liability by saying "it was the supplier's fault." You deploy it, you own the risk.

No community discussion yet for this question.