Anthropic warns Google search proposal will harm US AI investment

Anthropic, the AI company that has received $3 billion from Google, filed a legal challenge against a US government proposal for a rule that would ban Google from investing in AI startups on Friday, arguing that it would disrupt competition and cripple AI innovation in the country.

The proposal is part of the Justice Department’s antitrust case against Google’s parent company, Alphabet Inc., which a federal judge found guilty of illegally monopolizing online search and digital advertising markets.

Regulators now want to force Google to sell its Chrome web browser and prohibit it from funding or partnering with companies that manage search data, including AI companies like Anthropic.

Anthropic says Google ban will benefit its biggest rivals

In the court filing, Anthropic argued that cutting Google out of AI investments would create an “unjustified windfall” for larger competitors like OpenAI and Meta, the two most powerful players in artificial intelligence.

“A remedy that requires Google to terminate its relationship with Anthropic would harm both Anthropic and competition more generally,” Anthropic said in its filing.

It pointed out a glaring contradiction in the proposal: Google’s own AI division, DeepMind, would remain untouched. While regulators want to curb Google’s power, the rule doesn’t stop DeepMind from continuing to develop its own AI models, like Gemini, effectively leaving Google’s AI footprint intact while severing its financial ties to outside startups.

Anthropic is one of the few companies directly competing against OpenAI, which dominates the industry with ChatGPT. Anthropic’s own model, Claude, has been gaining traction, but training large-scale AI systems requires billions in funding.

The company has been aggressively raising capital to stay in the race, securing investments from Amazon alongside Google. But if Google is forced to pull out, Anthropic warned it could lose critical funding needed to expand computing power, train better models, and remain competitive.

The Federal Trade Commission (FTC) has already expressed concerns over the power that Big Tech companies have in AI. In a January report, the agency said that companies like Microsoft and Google often tie their AI investments to the use of their own cloud services and hardware, creating a closed-loop ecosystem that makes it even harder for smaller AI companies to compete.

Google executives address DeepSeek challenge and AI policy changes

While Google battles in court, its executives are also dealing with the new challenges from China’s AI industry. During an all-hands meeting in Paris on Wednesday, Google’s chief AI scientist, Demis Hassabis, told employees he was not concerned about the rise of DeepSeek, a Chinese AI startup that recently overtook OpenAI’s ChatGPT on Apple’s App Store rankings.

Hassabis dismissed claims that DeepSeek had trained its AI at a fraction of the cost of leading Western models, calling the numbers “exaggerated.” He added that the company had likely spent far more on hardware than it disclosed and relied on Western AI models to develop its own systems.

“We actually have more efficient, more performant models than DeepSeek,” Hassabis told employees. “So we’re very calm and confident in our strategy.”

But Hassabis did acknowledge DeepSeek’s potential, calling it “the best team I’ve seen come out of China.” He also flagged security and geopolitical risks, noting that several US agencies have already banned employees from using DeepSeek’s technology due to national security concerns.

Meanwhile, Google is also facing backlash over its decision to quietly change its AI policy. Employees grilled CEO Sundar Pichai on why the company removed a key section of its AI Principles, which previously banned the use of AI for weapons and surveillance.

Pichai passed the question to Kent Walker, Google’s president of global affairs, who revealed that Google executives had worked together to revise the rules in early 2024.

“Some of the strict prohibitions that were in v1 of the AI principles don’t jibe well with the more nuanced conversations we’re having now,” Walker explained, referring to the original 2018 guidelines that were implemented after public outrage over Google’s involvement in Project Maven, a controversial military AI project.

Walker defended the decision, telling the crowd in Paris that: “An awful lot has changed in those seven years.” He argued that as AI technology changes and grows, it is now being deployed in “far more nuanced scenarios” that don’t fit the original restrictions.

Cryptopolitan Academy: How to Write a Web3 Resume That Lands Interviews - FREE Cheat Sheet