Google and DeepMind must put aside years of rivalry and risk ruffling feathers to take on OpenAI, ex-employees and collaborators say

• Google and DeepMind must collaborate more despite the risk of internal turmoil, sources told the FT.
• A newly formed AI unit named Google DeepMind risks ruffling feathers, the publication reported.

Google and DeepMind are reportedly risking internal turmoil by collaborating more to thwart the growing threat posed by OpenAI.

The Financial Times reported the news, citing people familiar with the matter.

Google and DeepMind announced on April 20 that the London-based AI lab, bought for $500 million in 2014, would combine with Google's Brain division in an attempt to streamline AI resources and take on ChatGPT developer OpenAI.

But people familiar with the matter told The Financial Times that the restructuring was likely to "ruffle lots of feathers" and spark resignations, even if it meant giving Google a better shot at steering "a tighter ship" and navigating the challenges posed by competitors.

Former employees and collaborators at both companies told the paper that the newly formed unit, known as Google DeepMind, would need to put aside "years of rivalry" to see off the challenge from OpenAI.

Since the launch of ChatGPT in November, Google has been on high alert over the threat that rival generative AI tools pose. Microsoft's multibillion-dollar bet on the technology and integration into Bing could potentially siphon away users from Google's core search engine, for example.

But the integration of DeepMind with Google's AI lab is fraught with risk. Since Google's takeover almost a decade ago, DeepMind CEO Demis Hassabis has fought to ensure independence from the search giant, according to the Financial Times. This is so DeepMind can focus on its mission to achieve artificial general intelligence.

Insider previously reported that a DeepMind plan codenamed "Mario" was created to maintain independence, driven by fears that Google might misuse its technology.

But the existential threat posed by OpenAI has pushed Hassabis into a more central role at the heart of Google's efforts to keep pace with competitors during a period of intense competition.

DeepMind and Google have had a long history of collaboration across a number of products and projects, including an AI model that would allow humans to talk to machines, while working closely on AI ethics.

In an earnings call on Tuesday, Google CEO Sundar Pichai said the new unit would "help accelerate our progress and develop the most capable AI systems safely and responsibly."