American tech giant Meta Platforms is reportedly using an open-source artificial intelligence model developed by Alibaba Group Holding to reinvigorate its own stagnant AI efforts and aim for another Chinese AI win.
According to a Bloomberg report on Wednesday, Facebook owner Mehta used Alibaba's Qwen model, along with other open source models from Google and OpenAI, as part of the training process for a new model code-named “Avocado” that was scheduled to be released in the spring.
The report did not specify which Alibaba Qwen model was used.
Bloomberg reported that Meta's new model marks a departure from its previous strategy of open-sourcing its models, allowing users to access them only through official application programming interfaces (APIs). Meta has not officially announced a change in its AI strategy from an open source Llama model to a closed source revenue generation model.
Qwen is the world's most popular series of open source large-scale language models, and many US companies use it to develop their own applications. Brian Chesky, co-founder and CEO of San Francisco-based online hotel booking giant Airbnb, said the company is “heavily relying” on Alibaba's Qwen model to power its AI-driven customer service agents. Alibaba owns the South China Morning Post.
Alibaba shares fell 1.7% in Hong Kong on Thursday.
Meta's AI efforts have received a lot of attention since the tech giant ramped up spending over the summer to poach talent from rivals, including reports that the contract was worth $100 million.
