DeepSeek hired employees from Microsoft's AI labs in China
What's the story
DeepSeek, the Chinese AI start-up that recently sparked a $1 trillion selloff in US tech stocks, has a number of former Microsoft employees from the company's controversial AI labs in China.
At least four current DeepSeek employees, including a key department chief, had earlier worked at Microsoft Research Asia, the New York Post has reported.
The ex-Microsoft employees now at DeepSeek include the head of its AI "alignment team," which ensures that models follow a specific set of social values.
Career progression
DeepSeek's AI alignment team leader's journey
The leader of DeepSeek's AI alignment team spent a decade at Microsoft Research Asia (2013-2023).
He joined as a research intern and later became a senior researcher, working on large-scale language AI model training.
Another DeepSeek researcher also worked at Microsoft Research Asia, spending six years in the "natural language computing group" before joining the firm.
Both were "core contributors" on the research paper detailing DeepSeek's R1 reasoning model, which caused market turmoil last month.
Official statement
Microsoft's response to ties with DeepSeek researchers
Responding to the revelations, Microsoft acknowledged its ties with the researchers but played down the importance of their training at the company.
"Anyone who thinks that a handful of former Microsoft interns were the secret of DeepSeek's recent success doesn't understand what DeepSeek has accomplished," a Microsoft spokesperson said.
The company also said it makes full-time and internship candidates in China sign confidentiality and IP transfer agreements to prevent leaks.
Talent migration
Concerns over Microsoft's AI labs in China
Microsoft Research Asia has served as a breeding ground for some of China's best tech talent, many of whom later apply their skills at companies competing with US interests.
This has raised fears of intellectual property theft and talent poaching.
Critics say these labs are easy targets for IP theft (including AI secrets) and poaching of key talent.
However, there's no evidence of any DeepSeek employees with Microsoft ties doing anything wrong.
AI development
DeepSeek's AI chatbot raises questions and concerns
The advanced nature of DeepSeek's AI chatbot has raised questions about the start-up's claims that it was developed for less than $6 million and without NVIDIA's most advanced computer chips.
It has also sparked fears that US tech companies are at risk of losing their technological advantage to China.
Notably, the chatbot shows signs of censorship, and refuses to answer queries about China's leader Xi Jinping or the 1989 Tiananmen Square massacre.