Head of ChatGPT reveals deepfake fears and hints at new ‘more powerful’ AI

DEEPFAKE scams are “absolutely” something OpenAI is concerned about, according to the head of ChatGPT.

Peter Deng revealed these fears today before an audience at South by Southwest 2024 in Austin, Texas.

2

The head of ChatGPT was questioned about everything from AI deepfake scams to whether artists should be compensated when their work is used by the software.

Regarding the harm deepfakes could cause to society, Deng said this is “definitely something we are concerned about.”

The AI ​​expert urged caution when rolling out new software and believes the damage it could cause should be considered before it is released.

“The approach we have taken in releasing our technology is quite cautious,” he told an audience of more than 600 people.

He also hinted at more powerful versions of OpenAI’s models that won’t be released to the public.

“The models themselves are potentially more powerful than what we expose,” he explained.

Deng also emphasized that OpenAI is conscious and careful when building and developing products.

He expressed this caution in favor of accelerating the development of AI, despite other experts calling for a slowdown so that safety measures can be put in place.

“I’m somewhere in the middle. With any new technology there will be very positive use cases, but there are things we really need to take into account,” he said.

He thinks safe acceleration is possible if all developers are careful.

When it came to artists and whether they should be compensated if their work is used by AI, Deng had less to say.

Several members of the SXSW auditionee shouted “yes” when Deng was asked if performers deserved compensation.

The ChatGPT boss didn’t give a clear answer, but said artists needed to be part of the conversation.

He did reveal that he thinks AI can be a useful tool for artists.

“It’s going to accelerate the art world,” he said, explaining that it could help artists create work faster.