Smith urges tougher AI rules
1m
Microsoft President Brad Smith called for internationally coordinated, binding AI rules, including mandatory independent assessments of AI products. He also proposed kill switches and legal liability for harms caused by rogue AI agents, arguing that AI faces weaker oversight than household products.
Here are some more details from the proposal (doi.org/10.1007/s133...), designed to avoid this problem. Just as we may not want to permit an uninsured corporation (or one financially unable to meet its obligations) to take on high-stakes projects, the regulatory principle for AI should be similar.
October 12, 2026 at 3:50 PM
Here are some more details from the proposal (doi.org/10.1007/s133...), designed to avoid this problem. Just as we may not want to permit an uninsured corporation (or one financially unable to meet its obligations) to take on high-stakes projects, the regulatory principle for AI should be similar.
It may sometimes be appropriate to introduce regimes of strict liability for the operation of AI systems. The required insurance costs might then (ideally) incentivise greater AI safety. Of course, all this is very tricky. But there are lessons to be learnt from the case of corporate regulation. 3/3
October 12, 2026 at 2:14 PM
It may sometimes be appropriate to introduce regimes of strict liability for the operation of AI systems. The required insurance costs might then (ideally) incentivise greater AI safety. Of course, all this is very tricky. But there are lessons to be learnt from the case of corporate regulation. 3/3
again, we’re not regulating. we’re not doing even the basic due diligence to determine the impact of AI across all areas of public life incl war. the real near term risk to human life is here, not in some runaway agents and misalignment (which are bad too).
October 12, 2026 at 10:17 AM
again, we’re not regulating. we’re not doing even the basic due diligence to determine the impact of AI across all areas of public life incl war. the real near term risk to human life is here, not in some runaway agents and misalignment (which are bad too).
The basic proposal (albeit tricky): regulation should permit the use of autonomous AI in high-stakes settings only if structures are in place to ensure that those systems – or if infeasible, their legal representatives – bear responsibility (backed via proper insurance) for the systems’ actions. 2/3
October 12, 2026 at 2:14 PM
The basic proposal (albeit tricky): regulation should permit the use of autonomous AI in high-stakes settings only if structures are in place to ensure that those systems – or if infeasible, their legal representatives – bear responsibility (backed via proper insurance) for the systems’ actions. 2/3
by Martin McKee Reposted by Mike Daube
Who can say no? We’ve been thinking a lot about governance of AI, learning lessons from how society has responded to analogous challenges in the past. And we welcome feedback as this is really difficult
arxiv.org/abs/2610.13340
arxiv.org/abs/2610.13340
Who Can Say No? Control, Accountability and the Pace of Frontier AI
In September 2026, the leaders of the largest frontier AI laboratories said in public that the pace of capability improvement is too fast and that none of them can slow down alone, and agents under ev...
arxiv.org
October 12, 2026 at 8:19 AM
Who can say no? We’ve been thinking a lot about governance of AI, learning lessons from how society has responded to analogous challenges in the past. And we welcome feedback as this is really difficult
arxiv.org/abs/2610.13340
arxiv.org/abs/2610.13340
This NYT piece aptly raises the issue of *responsibility gaps* arising from the use of AI systems. In the paper below, I compare such responsibility gaps with those arising in the more widely studied case of corporate entities and discuss possible regulatory responses via liability arrangements.
Group Agency and Artificial Intelligence - Philosophy & Technology
The aim of this exploratory paper is to review an under-appreciated parallel between group agency and artificial intelligence. As both phenomena involve non-human goal-directed agents that can make a ...
link.springer.com
October 12, 2026 at 1:54 PM
This NYT piece aptly raises the issue of *responsibility gaps* arising from the use of AI systems. In the paper below, I compare such responsibility gaps with those arising in the more widely studied case of corporate entities and discuss possible regulatory responses via liability arrangements.
musk and all the AI oligarchs see the future now.
Uh oh, did someone’s website and personal account experience a big drop in activity after Ukraine hit a Russian data center?
October 12, 2026 at 10:26 AM
musk and all the AI oligarchs see the future now.
Reposted by Brigitte Nerlich
"Brad Smith said there needed to be global safety rules in place, including a legal obligation to tell a "victim" if their system is breached by a rogue AI agent as well as deciding when outside evaluators should sign off on an AI product before it is released." www.bbc.com/news/article...
AI is less regulated than milk, Microsoft president says
Brad Smith calls for global rules on AI, and says the UK is in the "best position" to lead the charge.
www.bbc.com
October 12, 2026 at 10:52 AM
"Brad Smith said there needed to be global safety rules in place, including a legal obligation to tell a "victim" if their system is breached by a rogue AI agent as well as deciding when outside evaluators should sign off on an AI product before it is released." www.bbc.com/news/article...
As *everything* becomes powered by AI, where do we even begin in distinguishing mil and non-mil uses of data centers. The commingling is a looming catastrophe.
October 12, 2026 at 10:10 AM
As *everything* becomes powered by AI, where do we even begin in distinguishing mil and non-mil uses of data centers. The commingling is a looming catastrophe.
I understand the claim "the rise of AI is also one of the biggest facts in the world."
Maybe there are 20 people whose conversations about AI are really critically important to that big fact. Maybe there are 200.
Either way, I'm sure they're all on a group chat already.
Maybe there are 20 people whose conversations about AI are really critically important to that big fact. Maybe there are 200.
Either way, I'm sure they're all on a group chat already.
October 12, 2026 at 3:46 AM
I understand the claim "the rise of AI is also one of the biggest facts in the world."
Maybe there are 20 people whose conversations about AI are really critically important to that big fact. Maybe there are 200.
Either way, I'm sure they're all on a group chat already.
Maybe there are 20 people whose conversations about AI are really critically important to that big fact. Maybe there are 200.
Either way, I'm sure they're all on a group chat already.