Elliott Thornley
@elliottthornley.bsky.social
Research Fellow at Oxford University's Global Priorities Institute.
Working on the philosophy of AI.
Working on the philosophy of AI.
Recent article on the POST-Agents Proposal!
Shutdownable Agents through POST-Agency — LessWrong
Summary
* Future artificial agents might resist shutdown.
* I present an idea – the POST-Agents Proposal – for ensuring that doesn’t happen.
* I p…
www.lesswrong.com
September 17, 2025 at 2:27 PM
Recent article on the POST-Agents Proposal!
Reposted by Elliott Thornley
A new working paper, "Shutdownable Agents through POST-Agency" by Elliott Thornley, is now available on our website. Read it here:
globalprioritiesinstitute.org/thornley-shu...
globalprioritiesinstitute.org/thornley-shu...
Shutdownable Agents through POST-Agency - Elliott Thornley
Many fear that future artificial agents will resist shutdown. I present an idea – the POST-Agents Proposal – for ensuring that doesn’t happen. I propose that we train agents to satisfy Preferences Onl...
globalprioritiesinstitute.org
June 2, 2025 at 1:43 PM
A new working paper, "Shutdownable Agents through POST-Agency" by Elliott Thornley, is now available on our website. Read it here:
globalprioritiesinstitute.org/thornley-shu...
globalprioritiesinstitute.org/thornley-shu...
'Where are you?' seems like a pretty normal question, but for 99.99% of human history it basically never made sense to ask it.
May 14, 2025 at 3:56 PM
'Where are you?' seems like a pretty normal question, but for 99.99% of human history it basically never made sense to ask it.
Daniel Kerson wrote a nice summary of a talk I gave in Singapore last month
kerson.ai/how-advanced...
kerson.ai/how-advanced...
How Advanced AI Agents Could Resist Shutdown and What Can Be Done - Kerson AI Solutions
I attended a talk today at the SASH (Singapore AI Safety Hub). The speaker today was Elliott Thornley who is a Research Fellow at Oxford University. Introduction As artificial intelligence advances, w...
kerson.ai
April 22, 2025 at 9:51 AM
Daniel Kerson wrote a nice summary of a talk I gave in Singapore last month
kerson.ai/how-advanced...
kerson.ai/how-advanced...
Our poster for TAIS 2025
April 8, 2025 at 6:55 AM
Our poster for TAIS 2025
A gif we made summarizing our 'Towards shutdownable agents' paper for TAIS 2025.
April 8, 2025 at 6:52 AM
A gif we made summarizing our 'Towards shutdownable agents' paper for TAIS 2025.
Gave a talk about the shutdown problem at the new Singapore AI Safety Hub!
March 24, 2025 at 6:01 PM
Gave a talk about the shutdown problem at the new Singapore AI Safety Hub!
I've got a new paper out open-access in AJP! It’s about critical-level and critical-range views in population axiology, and why I think they’re troubled by questions of identity between lives.
www.tandfonline.com/doi/full/10....
www.tandfonline.com/doi/full/10....
Critical-Set Views, Biographical Identity, and the Long Term
Critical-set views avoid the Repugnant Conclusion by subtracting some constant from the welfare score of each life in a population. These views are thus sensitive to facts about biographical identi...
www.tandfonline.com
March 13, 2025 at 9:17 AM
I've got a new paper out open-access in AJP! It’s about critical-level and critical-range views in population axiology, and why I think they’re troubled by questions of identity between lives.
www.tandfonline.com/doi/full/10....
www.tandfonline.com/doi/full/10....
Progress in AI has been rapid in recent years. By contrast, progress in 'opening sentences of papers about AI' has completely stalled.
December 4, 2024 at 4:02 PM
Progress in AI has been rapid in recent years. By contrast, progress in 'opening sentences of papers about AI' has completely stalled.
Reposted by Elliott Thornley
"Sure, the last 1000 grad students failed to solve the problem of induction, but that's no reason to think I can't do it."
November 26, 2024 at 9:16 PM
"Sure, the last 1000 grad students failed to solve the problem of induction, but that's no reason to think I can't do it."
Reposted by Elliott Thornley
We’re excited to announce our new research agendas – for philosophy, economics and psychology – have now been published! You can read them here: globalprioritiesinstitute.org/research-age...
Research agenda - Global Priorities Institute
The central focus of GPI is what we call ‘global priorities research’: research into issues that arise in response to the question, ‘What should we do with a given amount of limited resources if our a...
globalprioritiesinstitute.org
November 29, 2024 at 11:00 AM
We’re excited to announce our new research agendas – for philosophy, economics and psychology – have now been published! You can read them here: globalprioritiesinstitute.org/research-age...
[Pasting over an old Twitter thread about this post.]
The Shutdown Problem: Incomplete Preferences as a Solution — AI Alignment Forum
Preamble
This post is an updated explanation of the Incomplete Preferences Proposal (IPP): my proposed solution to the shutdown problem. The post is…
www.alignmentforum.org
November 26, 2024 at 10:55 AM
[Pasting over an old Twitter thread about this post.]
My favourite arguments against person-affecting views
1.
openairopensea.substack.com
November 19, 2024 at 3:40 PM
Paper! arxiv.org/pdf/2407.00805
With Alex Roman, Christos Ziakas, Leyton Ho, and Louis Thomson.
Quick thread explaining it.
With Alex Roman, Christos Ziakas, Leyton Ho, and Louis Thomson.
Quick thread explaining it.
arxiv.org
November 18, 2024 at 4:48 PM
Paper! arxiv.org/pdf/2407.00805
With Alex Roman, Christos Ziakas, Leyton Ho, and Louis Thomson.
Quick thread explaining it.
With Alex Roman, Christos Ziakas, Leyton Ho, and Louis Thomson.
Quick thread explaining it.
