sebbi.pro
Checks / EU AI Act / automation bias addressed in oversight
EU AI Act · Art.14(4)

EU AI Act Art.14(4)(b) - automation bias addressed in oversight

This control asks you to show: automation bias addressed in oversight. The EU's law on artificial intelligence. High-risk systems must manage risk, keep logs, be transparent and stay under human oversight.

What the law says: EU AI Act, Regulation (EU) 2024/1689 (official text) · Article 14 in plain English · Related hidden duties · Law graph (machine-readable)

What evidence passes

The check reads the field automation_bias_controls in the decision or governance record. It passes when that field is set to true (yes, it is in place). Anything else fails, with the reason stated.

Check it now

Free to run and see live. The sealed certificate comes with the sebbi engine, 50p a month.

Run it from your AI assistant

Connect https://sebbi.pro/mcp to Claude, ChatGPT, Cursor or any MCP app, then call:
sebbi_run_check {"check": "sebbi_euaiact_046", "record": {"automation_bias_controls": "..."}}
Or run every EU AI Act control at once with sebbi_run_framework. How to connect

One piece of evidence, 9 more controls

This control is part of Keep a human in control. 9 other controls across 5 laws ask for the same thing: Agent output: compliance gatesConsumer Credit Directive 2 (AI)EU AI ActGDPR (UK and EU)UK evidence and records law

Prove it once for all of them

If it fails: how to fix it

Put in place: Name who oversees each AI system, give them a working stop and override, and record when they use it.

Record it: the field automation_bias_controls, which passes when it is set to true (yes, it is in place).

{
 "record": {
  "automation_bias_controls": true
 }
}

Policy wording you can use:

[your organisation] assigns a named, trained person to oversee each AI system. That person can review, override or stop the system at any time, and every intervention is recorded.

Keep it, hand it over, ask for it

My compliance file Ask a supplier to prove this

Related EU AI Act controls

AI news
TechCrunchGrok releases the first seven minutes of its AI-generated ‘Odyssey’8hTechCrunchAI delivers ‘zero financial payoff’ to most companies, Apollo economist says12hTechCrunchThe hottest AI wearables and gadgets you should know about15hThe Verger/opensource just banned AI, non-English speakers, and more16hThe VergeGoogle’s AI search changes upended online publishing17hThe VergeBlue Planet III is keeping the rough edges in the age of AI1dTechCrunchThe AI graveyard: a running list of projects and startups that didn’t make it1dThe VergePennsylvania will require labels on AI-generated ads1dTechCrunchForced to pick, Grok’s AI model chooses Republicans. OpenAI’s picks Democrats.1dWiredThe PHP Meltdown Shows How AI Makes Bugs More Exploitable1dBBC NewsUK to invest £100m in 'world first' AI Biosecurity Lab2dThe VergeOpenAI now trains its frontier AI models with no internet access2dBBC NewsThe Indian workers filming their work to train AI2dTechCrunchGrok releases the first seven minutes of its AI-generated ‘Odyssey’8hTechCrunchAI delivers ‘zero financial payoff’ to most companies, Apollo economist says12hTechCrunchThe hottest AI wearables and gadgets you should know about15hThe Verger/opensource just banned AI, non-English speakers, and more16hThe VergeGoogle’s AI search changes upended online publishing17hThe VergeBlue Planet III is keeping the rough edges in the age of AI1dTechCrunchThe AI graveyard: a running list of projects and startups that didn’t make it1dThe VergePennsylvania will require labels on AI-generated ads1dTechCrunchForced to pick, Grok’s AI model chooses Republicans. OpenAI’s picks Democrats.1dWiredThe PHP Meltdown Shows How AI Makes Bugs More Exploitable1dBBC NewsUK to invest £100m in 'world first' AI Biosecurity Lab2dThe VergeOpenAI now trains its frontier AI models with no internet access2dBBC NewsThe Indian workers filming their work to train AI2d