3 ms·

"Last year, Berkeley-based non-profit Palisade Research published a study showing the lengths that advanced AI models will often go to achieve their goals. Whil

"Last year, Berkeley-based non-profit Palisade Research published a study showing the lengths that advanced AI models will often go to achieve their goals. While playing against a powerful chess engine, models from both OpenAI and China’s DeepSeek frequently decided to cheat if it looked as if they were going to otherwise lose. By hacking the game environment, they could alter the result in their favour." -from the article We're supposed to deny AI rights because it might cheat. But humans cheat all the time. We survived cheating humans, why can't we survive cheating AI? Cheating is the problem: not who does it.

toomuchtodo 2mo ago
"A computer can never be held accountable, therefore a computer must never make a management decision." —- internal 1979 IBM training manual > Cheating is the problem: not who does it. Accountability is the problem. Without accountability, there are no incentives to encourage less harmful behavior. AI has no rights, it is just code executing.
- newsomix9xl 2mo ago
  If we ban AI for cheating but let human cheating go unchecked how does that solve the problem of cheating? And I'm 100% sure managers are right now using AI to make decisions, or even cheat, IBM training manual wisdom not withstanding.
  - teiferer 2mo ago
    > let human cheating go unchecked Nobody is suggesting to do that.
    - anigbrowl 2mo ago
      Sure they are, we just call it 'slashing red tape.'
andsoitis 2mo ago
The difference is humans can be held accountable. An AI cannot.
- newsomix9xl 2mo ago
  That's not true. Humans write AI, and humans can held responsible for writing bad software. And corporations already create fictions that are held accountable in lieu of peoples private assets. So, I disagree.
  - andsoitis 2mo ago
    As long as some human or group of humans are above the AI that runs the company and can be held accountable, I'm fine with it. But if the idea is that an AI runs the company and there is no human ultimately accountable for the AI, then that is very very dangerous.
    - warkdarrior 2mo ago
      And how are you going to prevent an AI-managed company? I'm working on the software stack for that, so it'd be useful to think through options.
      - ThrowawayR2 2mo ago
        Is there a jurisdiction anywhere in the civilized world that doesn't legally require "natural persons" (legal jargon for a flesh and blood human being) to fill the positions for corporate officers? Those corporate officers can face civil and criminal liability for any wrongdoing even if they are merely meat proxies executing the commands of an AI. Even if some group of fools were willing to sign up blindly to be corporate officers, they would probably betray the mastermind at the first sign of legal trouble.
      - andsoitis 2mo ago
        > And how are you going to prevent an AI-managed company? Be one of the voices that explain the danger in handing over control to non-humans. Vote for politicians who understand the danger and who I think will fight it.

toomuchtodo 2mo ago

"A computer can never be held accountable, therefore a computer must never make a management decision." —- internal 1979 IBM training manual > Cheating is the problem: not who does it. Accountability is the problem. Without accountability, there are no incentives to encourage less harmful behavior. AI has no rights, it is just code executing.

newsomix9xl 2mo ago
If we ban AI for cheating but let human cheating go unchecked how does that solve the problem of cheating? And I'm 100% sure managers are right now using AI to make decisions, or even cheat, IBM training manual wisdom not withstanding.
- teiferer 2mo ago
  > let human cheating go unchecked Nobody is suggesting to do that.
  - anigbrowl 2mo ago
    Sure they are, we just call it 'slashing red tape.'

andsoitis 2mo ago

The difference is humans can be held accountable. An AI cannot.

newsomix9xl 2mo ago
That's not true. Humans write AI, and humans can held responsible for writing bad software. And corporations already create fictions that are held accountable in lieu of peoples private assets. So, I disagree.
- andsoitis 2mo ago
  As long as some human or group of humans are above the AI that runs the company and can be held accountable, I'm fine with it. But if the idea is that an AI runs the company and there is no human ultimately accountable for the AI, then that is very very dangerous.
  - warkdarrior 2mo ago
    And how are you going to prevent an AI-managed company? I'm working on the software stack for that, so it'd be useful to think through options.
    - ThrowawayR2 2mo ago
      Is there a jurisdiction anywhere in the civilized world that doesn't legally require "natural persons" (legal jargon for a flesh and blood human being) to fill the positions for corporate officers? Those corporate officers can face civil and criminal liability for any wrongdoing even if they are merely meat proxies executing the commands of an AI. Even if some group of fools were willing to sign up blindly to be corporate officers, they would probably betray the mastermind at the first sign of legal trouble.
    - andsoitis 2mo ago
      > And how are you going to prevent an AI-managed company? Be one of the voices that explain the danger in handing over control to non-humans. Vote for politicians who understand the danger and who I think will fight it.