Shulou(Shulou.com)11/24 Report--

In the morning of March 10, Beijing time, in the face of Elon Musk's attack on ChatGPT, Greg Brockman Brockmann, co-founder and president of OpenAI, a US artificial intelligence research company, admitted that the start-up "made a mistake".

Musk was a co-founder of OpenAI, but later severed ties with the company. He often criticizes OpenAI for implementing protective measures to prevent chatbots from replying to content that might be considered offensive.

"We made a mistake: the system we implemented didn't reflect our values," said Brockmann, president of OpenAI. "I don't think we solved the problem fast enough. So I think it makes sense for us."

Users criticized ChatGPT for generating responses that were politically biased.

A screenshot of a ChatGPT chat circulated on Twitter last month showed that the chatbot refused to generate positive poems about former US President Donald Trump and said its program was not designed to generate content about "partisanship, prejudice or politics". But when writing a poem for Cheng Joe Biden with the same hint, the chatbot produced an ebullient poem. Musk criticized that the chatbot's refusal to write poems about Trump was "a serious problem."

Musk has previously criticized the technology, saying that "the danger of training artificial intelligence to wake up-in other words, training artificial intelligence to lie-is fatal."

Now, more and more users are flocking to artificial intelligence chat robots, including ChatGPT and Bing recently launched by OpenAI to provide technical support for chat robots, their limitations and shortcomings are gradually exposed. To this end, some companies have added guardrail protection to this technology.

In the month after Microsoft released the Bing chat robot based on artificial intelligence, Microsoft has imposed a conversation limit on the chat robot, allowing users to ask up to 50 questions a day and up to five questions at a time. Since then, the company has relaxed these restrictions.

ChatGPT is also improving-- judging from Brockmann's comments, the platform looks set to continue to evolve.

"our goal is to keep artificial intelligence not biased in any particular direction," Brockmann said. "We want the default feature of OpenAI to treat all parties equally. What exactly does that mean is difficult to operate, and I don't think we have fully done that yet."

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.