news OpenAI Wants to Pay Millions to Prevent AI From Outsmarting Humans

[CashproofAi]

OpenAI Wants to Pay Millions to Prevent AI From Outsmarting Humans​

Artificial Intelligence company NDTV recently highlighted an unusual and highly paid job opening at OpenAI. The company behind ChatGPT is reportedly offering up to $445,000 annually for a role focused on preventing advanced AI systems from becoming dangerously autonomous. The position reflects growing concern inside the AI industry about how future systems could evolve faster than human oversight mechanisms can control.​

Why OpenAI Created This Role ? Know more.........​

 

[bmugford]

Well, it wasn't able to outsmart me when comparing lists of domains.

However, it still tried to gaslight me before finally admitting it was wrong.

https://twitter.com/x/status/2058632550453592510

 

[the_poet]

Well, it wasn't able to outsmart me when comparing lists of domains.

However, it still tried to gaslight me before finally admitting it was wrong.

https://twitter.com/x/status/2058632550453592510

Absolutely. AI is generally something you should only use if you already know exactly what you want to achieve and, in any case, you should always review the results. But if you dare to use this tool blindly, trusting the results without question, you're in for a rude awakening.

 

[CashproofAi]

Well, it wasn't able to outsmart me when comparing lists of domains.

However, it still tried to gaslight me before finally admitting it was wrong.

https://twitter.com/x/status/2058632550453592510

In numerous occasions, I have proved ChatGPT wrong. It is countless. Everyday. One day I can remember, rated 5 times "negative" at a stretch. Then I just closed the screen saying "Pl. stop". But for same work , the success rate of Claude was almost 90%. It is better not to say anything about Gemini, perplexity, Grok, Copilot etc, they are all under performers for my style of works. But they are all useful for particular task and all are improving everyday. My research on AI is still going on.

 

[CashproofAi]

Absolutely. AI is generally something you should only use if you already know exactly what you want to achieve and, in any case, you should always review the results. But if you dare to use this tool blindly, trusting the results without question, you're in for a rude awakening.

Thanks. That's the truth. You can't trust any AI platform blindly.