Users thank those sharing reports on DeepMind signing a classified DoD contract because they value the transparency around the company dropping its 2014 military-use ban.
Based on 4 visible X reactions from 6 accounts; directional sample.
Ask a question below.
Published answers will appear here.
@anderssandberg Thank you! Really appreciate your support 🙏 I'm particularly concerned that while some concerns about autonomous weapons will disappear as AI models get better/superhuman, concerns about (mass) surveillance and autonomous policing will only increase as the models improve
@BlackHC @erinkwoo Thank you for your brave act of transparency 🙏🏽 Truly 🫶🏽
@BlackHC @erinkwoo Thank you for sharing your story, I found it informative.
@davidmanheim Thank you! 🙏 Really appreciate it!
None of this is new. DeepMind's governance record: 2014: independent ethics board — reportedly a condition of the sale 2015: one informal meeting; then effectively abandoned 2018: AI Principles exclude weapons & surveillance 2025: exclusions dropped 2026: Pentagon contract https://x.com/BlackHC/status/2077009664794046875/photo/1
Internal critics warn informal culture cannot replace binding safety guardrails.
@anderssandberg Thank you! Really appreciate your support 🙏 I'm particularly concerned that while some concerns about autonomous weapons will disappear as AI models get better/superhuman, concerns about (mass) surveillance and autonomous policing will only increase as the models improve
None of this is new. DeepMind's governance record: 2014: independent ethics board — reportedly a condition of the sale 2015: one informal meeting; then effectively abandoned 2018: AI Principles exclude weapons & surveillance 2025: exclusions dropped 2026: Pentagon contract https://x.com/BlackHC/status/2077009664794046875/photo/1
Demis Hassabis then made a different bet: "Take the energy that was going into the trustless negotiation and put it into creating real trust." And: "Safety isn't about governance structures." Trust instead of governance. The Pentagon contract is the litmus test of that bet. https://x.com/BlackHC/status/2077009696154915048/photo/1
These are very important concerns, and it is heartening to see people at labs willing to speak out about these issues - but very disheartening that the labs won't stick to the values they claim. (It bodes particularly poorly for them managing the risks of future AI systems.) https://twitter.com/BlackHC/status/2077009476423647596
@scmallaby So far: silence. As far as I know, there was no company-wide announcement that the deal had been signed; Kent Walker's internal memo reportedly didn't even confirm it. DeepMind's safety culture is real. But safety culture is not governance. And trust is not governance.
What was signed? According to The Information, the contract permits "any lawful government purpose," requires Google to help adjust safety settings and filters at the government's request, and gives Google no veto over the government's lawful operational decisions.
DeepMind's founders saw the structural problem early. Sebastian Mallaby's @scmallaby new book "The Infinity Machine" describes Project Mario: years of negotiating for independence, a 3-3-3 board, even a $5B walk-away plan. Google resisted. The talks ended in 2021 with nothing.
Andreas is doing the right thing, and pointing at a key issue: AI supported mass surveillance can scale, will be used, and the only real protection is good governance - of AI and governments. https://twitter.com/BlackHC/status/2077009476423647596
Users thank those sharing reports on DeepMind signing a classified DoD contract because they value the transparency around the company dropping its 2014 military-use ban.
Based on 4 visible X reactions from 6 accounts; directional sample.
Ask a question below.
Published answers will appear here.
Demis Hassabis then made a different bet: "Take the energy that was going into the trustless negotiation and put it into creating real trust." And: "Safety isn't about governance structures." Trust instead of governance. The Pentagon contract is the litmus test of that bet. https://x.com/BlackHC/status/2077009696154915048/photo/1
These are very important concerns, and it is heartening to see people at labs willing to speak out about these issues - but very disheartening that the labs won't stick to the values they claim. (It bodes particularly poorly for them managing the risks of future AI systems.) https://twitter.com/BlackHC/status/2077009476423647596
@scmallaby So far: silence. As far as I know, there was no company-wide announcement that the deal had been signed; Kent Walker's internal memo reportedly didn't even confirm it. DeepMind's safety culture is real. But safety culture is not governance. And trust is not governance.
What was signed? According to The Information, the contract permits "any lawful government purpose," requires Google to help adjust safety settings and filters at the government's request, and gives Google no veto over the government's lawful operational decisions.
DeepMind's founders saw the structural problem early. Sebastian Mallaby's @scmallaby new book "The Infinity Machine" describes Project Mario: years of negotiating for independence, a 3-3-3 board, even a $5B walk-away plan. Google resisted. The talks ended in 2021 with nothing.