Stay informed with free updates
Simply sign up to the Artificial intelligence myFT Digest — delivered directly to your inbox.
Meta and Elon Musk’s xAI have been vying for a partnership with chatbot-maker Character.ai, as deep-pocketed tech groups swarm around the hottest artificial intelligence start-ups in Silicon Valley.
Facebook owner Meta recently held early discussions over a tie-up with Character.ai, which uses large language models to generate conversation in the style of various figures and personas, according to four people familiar with the matter.
The groups discussed their top researchers working closely together on initiatives such as pre-training and developing models, some of the people said.
The fledgling group, which is backed by Silicon Valley venture firm Andreessen Horowitz, has also held exploratory talks with xAI over a similar partnership, said two people with knowledge of the discussions.
However, the talks have not resulted in any deal being struck, the people added.
The interest in Character.ai is the latest sign of how leading tech groups are seeking to secure partnerships and investments in top AI start-ups, as they race one another to develop the cutting-edge technology.
Meta said in September that it was incorporating AI persona chatbots across Instagram, Facebook and WhatsApp, including several that take on celebrity characters such as rapper Snoop Dogg.
Meanwhile, xAI has developed its own chatbot, Grok, which is available to premium subscribers on X, the social media platform that Musk also owns.
The discussions were focused on advancing research rather than an acquisition, one of the people said.
Big Tech groups have been wary of attempting full takeovers of AI start-ups over fears of regulatory action around the world.
Microsoft’s $13bn alliance with OpenAI is being examined by UK and US competition authorities, despite the two groups insisting their partnership is not a merger.
Meta has also made contact with other AI companies about potential partnerships, said a person with knowledge of its strategy.
Such moves come as the platform ramps up its investments in AI and sets its sights on becoming “the leading AI company in the world”, according to chief executive Mark Zuckerberg.
Adept, an AI agent start-up run by former OpenAI and Google AI developers, has also held talks with Meta about a sale or strategic partnership, according to one person familiar with the matter. News of the Adept deal was first reported by The Information on Wednesday.
Meta has taken a different approach to its biggest rivals in developing AI models.
The company’s flagship Llama models are open-source, meaning that their source code is freely available to the public and developers building on top of the platform. OpenAI, Anthropic and Google, by contrast, have developed “closed” proprietary models.
Meta’s models are a cheaper alternative for enterprises and developers but investors have baulked at the cost of keeping pace with rivals.
Last month, Meta lost nearly $200bn in market value when Zuckerberg vowed to increase AI spending as part of its work towards developing artificial general intelligence — loosely defined as AI capable of outperforming humans across a range of tasks.
Musk’s xAI is closing in on a $6bn fundraise, which the serial entrepreneur hopes will help it gain ground on OpenAI, Google and Meta.
He has spoken about his desire to build “maximum truth-seeking AI” and his pitch to investors is that access to data and expertise from his other companies — such as Tesla and X — give him an edge over rivals.
Noam Shazeer, Character.ai’s founder and a former Google researcher, was one of the authors of a 2017 paper that proposed the transformer model, which underpins the best AI models today.
According to one person who knows Shazeer, he is focused on building AGI, and seeking out more resources to do so. Character.ai is also exploring partnerships with other groups, said one person familiar with its strategy.
Character.ai and Meta declined to comment. Musk did not respond to a request for comment.
Eugen Boglaru is an AI aficionado covering the fascinating and rapidly advancing field of Artificial Intelligence. From machine learning breakthroughs to ethical considerations, Eugen provides readers with a deep dive into the world of AI, demystifying complex concepts and exploring the transformative impact of intelligent technologies.
