AI warning from maker of ChatGPT: 'We might need to slow it down'
"We have to be careful with AI. GPT-4 performs at a human level on a variety of professional and academic benchmarks. I'm worried that these models could be used specifically for large-scale disinformation. Now that they've gotten better at writing computer code, they can be used for offensive cyber attacks," OpenAI CEO Sam Altman warned.
- Tech
- A News
- Published Date: 10:24 | 20 March 2023
- Modified Date: 10:30 | 20 March 2023
OpenAI CEO Sam Altman warned that artificial intelligence has the potential to replace employees, spread "disinformation" and enable cyberattacks.
Altman spoke to ABC News two days after his company unveiled its latest language model, called GPT-4.
"We have to be careful with AI. GPT-4 performs at a human level on a variety of professional and academic benchmarks. I'm worried that these models could be used specifically for large-scale disinformation. Now that they've gotten better at writing computer code, they can be used for offensive cyber attacks," he said.
Altman said his company is working to set "safety limits" on artificial intelligence, adding, "I think people should be happy that we're a little afraid of it."
Currently, GPT-4 is available to a limited number of users on a trial basis.
Initial reports suggest the model is significantly more powerful and potentially more dangerous than its predecessor.