Prohibited AI Practices
AI uses that are completely illegal because they violate basic human rights, like government social scoring systems or real-time facial recognition tracking in public spaces. Just as certain dangerous chemicals are banned from consumer products, these AI practices are banned outright, with no compliance "workaround."
The Simple Version
AI uses that are completely illegal because they violate basic human rights, like government social scoring systems or real-time facial recognition tracking in public spaces. Just as certain dangerous chemicals are banned from consumer products, these AI practices are banned outright, with no compliance "workaround."
Detailed Explanation
Under regulatory frameworks like the EU AI Act (Article 5), prohibited practices include AI systems that deploy subliminal techniques to materially distort human behavior, exploit vulnerabilities of specific groups (e.g., children or the elderly), conduct social scoring by public authorities, or perform real-time remote biometric identification in publicly accessible spaces (with very narrow, strictly defined law enforcement exceptions). These are classified as "Unacceptable Risk."
Key Characteristics
- Absolute Ban: No conformity assessment or mitigation can make these practices legal.
- Extraterritorial Scope: Applies to any organization offering these systems within the regulated jurisdiction, regardless of where the company is based.
- Severe Penalties: Violations can result in massive administrative fines (e.g., up to €35 million or 7% of global annual turnover under the EU AI Act).
Why It Matters
Organizations must screen all proposed AI use cases against prohibited lists during the earliest ideation phase. Attempting to develop or deploy a prohibited practice is a hard stop that can result in immediate project termination, reputational damage, and severe legal liability.
Common Misconceptions
- Myth: If we get explicit user consent, we can use real-time facial recognition in public spaces.
- Myth: We can just tweak the algorithm slightly to make a banned use case compliant.