sebbi Oversight ProofProof a real person checked the AI

Prove a real person checked the AI

The law says a human must really check high-risk AI decisions, not rubber-stamp them. sebbi proves it, review by review: when it happened, how long they spent, whether they ever disagree, and (if you switch it on) that a person was at the screen.

Not made in advance: every decision carries a live Bitcoin code
Not edited afterwards: sealed into a chain nobody can change
Real time spent: under 5 seconds is flagged
Real disagreement: a reviewer who never overrules the AI is flagged
A person at the screen: optional camera check, head in the frame or computer says no

Free for 3 months on one device (up to 100 a month). Then 50p per device per month.

sebbi app · sebbi Teams · Privacy · Terms

AI news

TechCrunchTrajectory raises $45M to make AI model training accessible to nontechnical experts42m TechCrunchAnthropic warns that it’s very close to RSI capability, aka self-building AI1h The VergeAnthropic clamps down on Claude’s AI video game clones5h Ars TechnicaClaude’s new rules against game cloning won’t stop the AI copying boom5h The VergeOpenAI pitches data center power plants as a way to protect nearby areas5h The VergeThe New York Times sues Brave web browser for AI scraping and bypassing paywalls6h MIT Technology ReviewRaising the quality of insights in agentic analytics6h TechCrunchElon Musk throttles Grok Bot limits due to ‘unsustainably high’ costs7h TechCrunchOmen AI adds liquid know-how in $16 million Aquatic Labs acquisition8h TechCrunchOpenAI is finally merging ChatGPT’s Chat and Work tabs9h WiredWhen You Should—and Shouldn’t—Use AI to Search the Web11h WiredHundreds of Mundane AI Apps Are Hiding Sexually Abusive Deepfake Features13h MIT Technology ReviewThe agentic web is here14h WiredChatbots Are Leading Hikers Astray, With Dangerous Consequences1d BBC NewsThese firms you've never heard of are feeding fears the AI bubble could burst1d The VergeI built an AI Tamagotchi before Meta could2d WiredSpooky New Infostealer Malware Uses AI to ‘Autodox’ Victims2d WiredAn AI Agent Injected Drugs Into a Live Mouse’s Brain2d TechCrunchTrajectory raises $45M to make AI model training accessible to nontechnical experts42m TechCrunchAnthropic warns that it’s very close to RSI capability, aka self-building AI1h The VergeAnthropic clamps down on Claude’s AI video game clones5h Ars TechnicaClaude’s new rules against game cloning won’t stop the AI copying boom5h The VergeOpenAI pitches data center power plants as a way to protect nearby areas5h The VergeThe New York Times sues Brave web browser for AI scraping and bypassing paywalls6h MIT Technology ReviewRaising the quality of insights in agentic analytics6h TechCrunchElon Musk throttles Grok Bot limits due to ‘unsustainably high’ costs7h TechCrunchOmen AI adds liquid know-how in $16 million Aquatic Labs acquisition8h TechCrunchOpenAI is finally merging ChatGPT’s Chat and Work tabs9h WiredWhen You Should—and Shouldn’t—Use AI to Search the Web11h WiredHundreds of Mundane AI Apps Are Hiding Sexually Abusive Deepfake Features13h MIT Technology ReviewThe agentic web is here14h WiredChatbots Are Leading Hikers Astray, With Dangerous Consequences1d BBC NewsThese firms you've never heard of are feeding fears the AI bubble could burst1d The VergeI built an AI Tamagotchi before Meta could2d WiredSpooky New Infostealer Malware Uses AI to ‘Autodox’ Victims2d WiredAn AI Agent Injected Drugs Into a Live Mouse’s Brain2d