When you buy through affiliate links in our content, we may earn a commission at no extra cost to you. Learn 
Google is ramping up efforts to create AI that mirrors human reasoning, drawing parallels to OpenAI’s recent breakthroughs.
This development introduces a fresh chapter in the ongoing competition between the tech giant and the swiftly expanding AI startup.
According to Bloomberg, over the past few months, several teams within Alphabet Inc.’s Google have made significant headway in AI reasoning capabilities. Per the insiders who requested anonymity due to the private nature of the information, this advanced AI software is designed to excel in solving complex, multi-step problems, particularly in fields such as mathematics and computer programming.
Google vs. OpenAI
AI researchers are racing to develop reasoning models that represent the next major leap in artificial intelligence. Google, similar to OpenAI, is focused on mimicking human-like reasoning using a process known as chain-of-thought prompting. This method, pioneered by Google, involves the AI briefly pausing before responding to a query while it processes multiple related questions behind the scenes, ultimately offering a comprehensive response.
While Google declined to comment on these efforts, the company is clearly determined to maintain its position in the AI race. Since OpenAI launched its popular ChatGPT, Google has taken steps to regain ground, merging its top research teams into Google DeepMind and enhancing collaboration between its research and product divisions.
Nevertheless, Google has been cautious in rolling out new AI products, taking time to consider ethical concerns and ensuring that public trust in its brand is upheld.
Google’s AI Innovations and the Future of the Industry
OpenAI’s introduction of its o1 model, internally referred to as Strawberry, in mid-September sparked some concern within Google’s DeepMind division about falling behind. However, with Google’s recent AI developments, those concerns have lessened, according to another insider.
Despite the slower pace of product releases, Google remains a formidable force in AI, says Oren Etzioni, a well-respected AI researcher and founder of TrueMedia.org, a nonprofit dedicated to combating political disinformation. Etzioni noted that Google has always had top-tier capabilities but has taken a more cautious approach to releasing its technology. “It’s a marathon, and anyone can come out on top,” he said.
In July, Google unveiled AlphaProof, a math reasoning tool, and AlphaGeometry 2, an updated model designed to tackle geometric problems. These systems solved four out of six problems from the prestigious International Mathematical Olympiad, an annual contest focusing on algebra and geometry, Google shared in a blog post.
At its May developer conference, Google previewed an AI assistant named Astra, which can use a phone’s camera to interpret its surroundings and answer real-world questions, such as locating misplaced items. Google indicated that some of Astra’s features may be integrated into its flagship AI model, Gemini, by year’s end.
“Advanced mathematical reasoning is essential for modern AI,” said Demis Hassabis, CEO of Google DeepMind, in a July post on X.
Advanced mathematical reasoning is a critical capability for modern AI. Today we announce a major milestone in a longstanding grand challenge: our hybrid AI system attained the equivalent of a silver medal at this year’s International Math Olympiad! https://t.co/O1QvMC02uM
— Demis Hassabis (@demishassabis) July 25, 2024
