Google insists it will maintain AI lead over China’s DeepSeek

Demis Hassabis, Google’s AI Chief, feels confident in the company’s ability to maintain its lead in the AI sector despite the emergence of China’s DeepSeek. Hassabis claimed Google has models that can not only compete but beat out the field.

DeepSeek, the infamous low-cost AI model created by a nondescript Chinese group, has had the US’s top AI startups and companies worried and on the defensive since its emergence and consequent run to the top of Apple and Android app stores.

In fact, investors dumped American tech stocks, including Nvidia, Vertiv Holdings, and Microsoft, based on fears that the existence of models like DeepSeek could threaten the dominance giants in the sector currently enjoy. 

Google sounded confident during its all-hands meeting in Paris 

Google’s AI chief told employees at an all-hands meeting in Paris on Wednesday that he’s not worried about DeepSeek, insisting to his company’s staff that it possesses superior artificial intelligence technology. 

During the meeting, Alphabet CEO Sundar Pichai read aloud a question about DeepSeek that wanted to know “what lessons and implications” Google has been able to glean from DeepSeek’s success as the company trains future models. The question was an AI summary of submissions from employees.

DeepSeek has claimed, according to a research paper released last month, that its AI model was trained at a fraction of the cost of other leading models like OpenAI’s funds-guzzling ChatGPT. 

Google DeepMind CEO Demis Hassabis provided the answer. In his response, he urged people to look at the details, which suggests an exaggeration on DeepSeek’s end. 

Hassabis surmised that DeepSeek’s reported cost of AI training was likely “only a tiny fraction” of the total cost of developing its systems. He said DeepSeek most likely used a lot more hardware than it revealed and that it also relied on Western AI models.

“We actually have more efficient, more performant models than DeepSeek,” Hassabis boasted, adding that Google is quite “calm and confident” that it has all the ingredients it needs to maintain leadership. 

While Hassabis boasted about Google’s capabilities, he also acknowledged DeepSeek’s accomplishments and how impressive they are. He called the lab responsible for building the model possibly “the best team” out of China and something to be taken seriously.

While Hassabis acknowledged DeepSeek’s impressive achievements, he also noted the security and geopolitical implications of their advancements, as several US agencies have restricted staff from using DeepSeek, citing security concerns.

Employee concerns were also addressed during the meeting 

During the all-hands meeting, Google executives also took the opportunity to address employee concerns regarding the recent changes to the company’s AI principles, particularly the removal of a pledge against using AI for weapons or surveillance.

Pichai read aloud an AI-summarized version of the questions, ending with “Why did we remove this section?”

The question was directed to Kent Walker, Google’s president of global affairs, who has reportedly worked with Hassabis, James Manyika, a senior vice president at the company, and others on the effort that “shifted our approach,” starting last year.

The pledge removal affected AI principles established in 2018 after Google refused to renew a government contract called Project Maven, which helped to analyze and interpret drone videos using AI.

Walker explained that the shift in AI principles reflects the evolving nature of the technology and the need for more nuanced discussions compared to the strict prohibitions that governed it in 2018.

He also pointed out that the tech is now being used in more diverse scenarios, though it is still unclear why that should cause a removal of Google’s pledge not to use AI for surveillance or weapons. 

Cryptopolitan Academy: Are You Making These Web3 Resume Mistakes? - Find Out Here