Musk aims to create a super AI to rival ChatGPT

Elon Musk is reportedly meeting with top-level experts who previously worked at Google to develop a new AI system to rival ChatGPT.

The Information cites three sources familiar with the matter, stating that the American billionaire has approached several AI researchers in recent weeks to discuss establishing a new laboratory aimed at developing a replacement for ChatGPT. Among them is Igor Babuschkin, a former senior expert at Google DeepMind and OpenAI.

While Babuschkin confirmed his departure from DeepMind last week, he has not officially joined Musk’s team. He stated that he “wants to work with Musk” to create something in the field of LLM (large language model) technology, which is behind the success of ChatGPT. However, he added that creating an uncontrolled chatbot is not Musk’s goal.

Billionaire Elon Musk is reported to be meeting with high-level experts who previously worked at Google to develop a new AI system to confront ChatGPT. The Information cited three insiders who said that in recent weeks, the American tycoon has been in contact with some artificial intelligence researchers to discuss establishing a new laboratory to develop a solution to replace ChatGPT. Among them is Igor Babuschkin, a former senior expert of Google DeepMind and OpenAI.

Babuschkin confirmed leaving DeepMind last week, but has not officially joined Musk’s team. He said he “wanted to work with Musk” to create something in the LLM (large language model) space, the technology behind ChatGPT’s success. However, Babuschkin said that creating an uncontrolled chatbot is not Musk’s goal.

Elon Musk co-founded OpenAI in 2015 with other famous figures such as Sam Altman, Reid Hoffman, and Peter Thiel. Initially, it was a non-profit open-source organization that focused on developing AI to counter Google, which they believed did not pay enough attention to AI safety.

Musk resigned as chairman of the company in February 2018. In 2020, he expressed dissatisfaction when OpenAI licensed Microsoft to exclusively exploit GPT-3. “This is contrary to the original open mission. OpenAI has basically been taken over by Microsoft,” Musk commented on Twitter at the time.

Recently, he also criticized OpenAI for not taking measures to prevent ChatGPT from creating conversations containing false information or offending users. In December 2022, he posted a picture on Twitter of ChatGPT refusing to discuss support for fossil fuels. Musk said, “There is a significant risk in training AI to lie.” Later, he said, “ChatGPT is good to the point of being scary.”

ChatGPT is considered one of the major technological breakthroughs in the past year. After its launch, it heated up the AI race, forcing tech giants to take action. From Silicon Valley giants to Asian startups, they all declared their participation in the super AI wave.

Despite the allure of ChatGPT, experts have also warned of the potential risks associated with this AI. One of the concerns is the ability to generate responses independently. Sam Altman, CEO of OpenAI, also said that the output of ChatGPT actually contains many errors. “It knows a lot, but the dangerous thing is that it is confident and significantly wrong,” he wrote on Twitter in early December 2022.