Robot Constitution
The Robot Constitution is a security ruleset part of AutoRT set by Google DeepMind in January 2024 for its AI products. The rules are inspired by Asimov's Three Laws of Robotics.The rules are applied to the underlying large language models of the helper robots.{{Cite web|url=https://www.theverge.com/2024/1/4/24025535/google-ai-robot-constitution-autort-deepmind-three-laws|title=Google wrote a "Robot Constitution" to make sure its new AI droids won't kill us|first=Amrita|last=Khalid|date=January 4, 2024|website=The Verge}}{{Cite web|url=https://www.businessinsider.com/google-deepmind-rules-ai-robots-safer-in-your-home-2024-1|title=Google DeepMind has new rules to make sure AI robots behave when tidying your home|first=Hasan|last=Chowdhury|website=Business Insider}}{{Cite web|url=https://www.siliconrepublic.com/machines/google-deepmind-robot-constitution-real-world-safety|title=DeepMind is training robots for real-world activities|first=Leigh Mc|last=Gowran|date=January 5, 2024|website=Silicon Republic}}{{Cite web|url=https://www.tomsguide.com/news/google-is-using-ai-to-teach-robots-household-chores-heres-the-result|title=Google's DeepMind is using AI to teach robots household chores — here's the result|author1=Ryan Morrison|date=January 5, 2024|website=Tom's Guide}}{{Cite web|url=https://uk.pcmag.com/ai/150336/google-taps-asimovs-three-laws-of-robotics-for-real-robot-safety|title=Google Taps Asimov's Three Laws of Robotics for Real Robot Safety|date=January 4, 2024|website=PCMag UK}}{{cite web | url=https://techcrunch.com/2024/01/04/google-outlines-new-methods-for-training-robots-with-video-and-large-language-models/ | title=Google outlines new methods for training robots with video and large language models | date=4 January 2024 }}
Rule number 1 is a robot “may not injure a human being” .
References
{{Reflist}}
Category:Existential risk from artificial general intelligence
{{Google AI}}
{{ai-stub}}
{{Google-stub}}