Fully autonomous AI agents should not be developed

https://huggingface.co/papers/2502.02649

by eamag • 1 year ago

38 44 1 year ago

lysace• 1 year ago

Our analysis reveals that risks to people increase with the autonomy of a system: The more control a user cedes to an AI agent, the more risks to people arise. Particularly concerning are safety risks, which affect human life and impact further values.

upghost2• 1 year ago

This is a purely procedural question, not supporting or critiquing in any way-- other than this reads kind of like an editorial with the format of a scientific paper. The question is... are there rules about what constitutes a paper or can you just put whatever you want in there as long as you follow "scientific paper format"?

dhruvbatra• 1 year ago

This looks like ICML formatting (and the submission deadline just passed).

ICML25 has an explicit call for position papers: https://icml.cc/Conferences/2025/CallForPositionPapers

upghost2• 1 year ago

Wow, great observation. Thank you. Makes sense. I'd never heard of a "position paper" before.

bamboozled2• 1 year ago

People are going to be developing these no matter what. Whether it wipes us out or not is just up to fate really.

esafak3• 1 year ago

We can constrain their use, as with nuclear materials.

fizx• 1 year ago

Nuclear materials have the advantages of being rare, dangerous to handle, and hard to copy over the internet.

johanneskanybal• 1 year ago

No not really. There's no power in the world that can restrain this in it's current form even mildly much less absolutly. Why do you think that would be even slightly possible?

redeux2• 1 year ago

At some point in the probably near future it will be much simpler to create an autonomous AI agent than a nuclear bomb.

roenxi• 1 year ago

Despite doing a pretty decent job of containing the risk we're still on the clock until something terrible happens with nuclear war. Humanity appears to be well on track to killing millions to billions of people; rolling the dice relatively regularly waiting for a 1% chance to materialize.

If we only handle AI that well doom is probable. It has economic uses, unlike nuclear weapons, so there will be a thriving black market dodging the safety concerns.

gcanyon• 1 year ago

It is a lot easier to detect illicit nuclear work compared to illicit AI work.

bamboozled2• 1 year ago

Look at who has access to US nuclear codes now. I don’t believe it’s as constrained as you think.

ASalazarMX2• 1 year ago

In the incredible case that we develop fully autonomous agents capable of crippling the world, that would mean we developed fully autonomous agents capable of keeping it safe.

Unless the first one is so advanced no other can challenge it, that is.

wendyshu• 1 year ago

Fallacious

grayfaced3• 1 year ago

How did you jump to that conclusion? The agent will be limited by the capabilities under its control. We have the technological ability to cripple world now and we don't have the technological means to prevent it. Give one AI control of the whole US arsenal and the objective of ending the world. Give another AI the capabilities of the rest of the world and the objective of protecting it. Would you feel safe?

bdangubic• 1 year ago

the best way to get people to stop doing X is to tell them not to do X. works so well with my kid :)

mark_l_watson• 1 year ago

I really enjoy Margaret Mitchell‘s podcast (she is the first author on the paper), and perhaps I missed something important in the paper, but:

Shouldn’t we treat separately autonomous agent we write ourselves, or purchase to run on our own computers, on our own data and that use public APIs for data?

If Margaret is reading this thread, I am curious what her opinion is.

For autonomous agents controlled by corporations and governments, I mostly agree with the paper.

in3d• 1 year ago

I'd recommend looking for other sources of information if you're relying on someone who co-authored the paper that introduced the most misleading and uninformed term of the LLM era: "stochastic parrot".

currymj• 1 year ago

it was a pretty defensible term at the time the paper came out, in the context of how LLMs were being trained and used.

in this paper, it's clear that the authors don't think modern LLM-based systems are just stochastic parrots.

artninja1988• 1 year ago

This is quite concerning seeing that the authors are all affiliated with huggingface. Hopefully they won't start censoring what models you can or can't upload because they seem certain things shouldn't be developed.

redeux2• 1 year ago

The paper described a level 5 fully autonomous agent as one that can:

create code(user request);

execute();

Is this not possible with tool use alone, so long as the agent has access to a tool that can execute arbitrary code?

jcarrano• 1 year ago

I feel that these kinds of statements are more effective at promoting AI than limiting it. It reinforces the assumption that such powerful AI is behind the corner. It hypes up AI and the result is likely more money and resources being put into it.

Imagine if the A-bomb was being openly developed. What title would have contributed more to funding and research, "The A-bomb (is terribly powerful and) should not be developed" or "The A-bomb will never work"? Except the A-bomb did work and in a surprisingly short time, while autonomous AGI is still a conjecture.

rtcode_io• 1 year ago

Yet, we all know we will!

asdasdsddd• 1 year ago

This has to be the least interesting paper I've ever read with the most surface level thinking.

• Simple→Tool Call: Inaccuracy propagated to inappropriate tool selection.
• Multi-step: Cascading errors compound risk of inaccurate or irrelevant outcomes.
• Fully Autonomous: Unbounded inaccuracies may create outcomes wholly unaligned with human goals.

Just... lol

geor9e2• 1 year ago

Too late. I added a 5 minute cron job for cursor AI's compose tab in agent mode that keeps replying "keep going, think of more fixes and features, random ideas as fine, do it all for me". I won't pull the plug.

voisin• 1 year ago

You’ve created a monster

rickydroll• 1 year ago

you say monster, I say plastic pal who's fun to be with

djohnston• 1 year ago

How do you programmatically interact w cursor??

geor9e2• 1 year ago

Same way I botted runescape as a child, by simulating user inputs with any macro app.

fragmede• 1 year ago

AGI confirmed.

Animats• 1 year ago

Most of the arguments presented also apply to corporations.

There's no mention of externalities. That is, are the costs of AI errors borne by the operator of the AI, or a third party.

numba888• 1 year ago

Hmm.. agent cannot do self-supervised learning without actually doing it. The trick is to keep it in a sandbox.

satisfice• 1 year ago

No one should be allowed to develop software that has bugs in it that lead to unlawful harm to others. And if they do it anyway they should be punished lawfully.

The thing with autonomous AI is that we already know it cannot be made safe in a way that satisfies lawmakers who are fully informed about how it works… unless they are bribed, I suppose.

Fully autonomous AI agents should not be developed

Related Stories