Sam Altman suggests that Open AI still doesn't fully understand how GPT works, despite rapid growth

Sam Altman

OpenAI has raised tens of billions of dollars to develop artificial intelligence (AI) technologies that are changing the world, and its ChatGPT chatbot uses large language models (LLMs) to create humanlike conversational dialogue. However, Open AI’s CEO, Sam Altman while addressing the International Telecommunication Union (ITU) AI for Good Global Summit in Geneva, Switzerland, acknowledged that the company still does not know how its LLMs work.

"We certainly have not solved interpretability," Altman suggested, as quoted by the Observer, essentially suggesting that the company has yet to figure out how to trace back its AI models. On being further probed whether they should release new and more powerful models, he said, “these systems [are] generally considered safe and robust.”

Additionally, Altman added that they don’t understand “what’s happening in your brain at a neuron-by-neuron level,” and they follow some standard rules to understand how a human mind would like to explain any particular query.

Altman’s unsatisfying answer highlights a real issue in the AI space going ahead, as many experts still struggle to explain the autonomous answers that AI chatbots provide to any query that's being thrown at them, raising safety concerns or privacy issues.

However, he somewhat assured GPT users that, “It does seem to me that the more we can understand what’s happening in these models, the better. I think that can be part of this cohesive package to how we can make and verify safety claims.”

Sam Altman also expressed his enthusiasm on social media platform X, after ChatGPT’s integration into Apple devices was confirmed at the ongoing WWDC 2024 event. He stated, “Very happy to be partnering with Apple to integrate ChatGPT into their devices later this year! Think you will really like it.”