post-thumb

Google AI Chief assures employees of the company's AI leadership over DeepSeek

At the Mobile World Congress in Barcelona, Google DeepMind CEO Demis Hassabis addressed concerns about China's emerging AI company, DeepSeek. During a recent all-hands meeting in Paris, Sundar Pichai, CEO of Alphabet, highlighted employee inquiries regarding DeepSeek's rapid market success, particularly its rise to the top of the App Store, surpassing ChatGPT.

Hassabis expressed confidence in Google's AI capabilities, asserting that some of DeepSeek's claims regarding the efficiency and cost-effectiveness of its AI training are overstated. He suggested that the actual costs involved in DeepSeek's development are likely much higher than reported, as the company may have utilized more hardware and relied on existing Western AI models. Despite this, Hassabis acknowledged the talent within DeepSeek, remarking it features "the best team" emerging from China, which presents security and geopolitical considerations, especially given the restrictions placed by U.S. agencies on the use of DeepSeek's technology.

In addition to discussing competitor concerns, Google executives responded to employee questions about changes to the company's AI Principles, specifically the removal of a commitment against using AI for military or surveillance purposes. Kent Walker, Google's president of global affairs, explained that the evolving nature of AI technology necessitated a more nuanced approach to these principles, which were originally established in 2018 following the controversy surrounding Project Maven, a military contract involving AI.

Overall, the discussions at the conference reflect Google's strategic positioning in the AI landscape amid growing competition and shifting ethical considerations.

Share: