23.1 C
New York

Google ready to compete with OpenAI and its new AI model that reasons like humans

The AI arms race is heating up as Google is said to be working on a new reasoning model, following OpenAI’s lead. According to a Bloomberg report, multiple Google teams are building an AI that can do human-like reasoning, this is a direct response to OpenAI’s O1 model (formerly known as “Strawberry”.) But what sets Google apart is a technique called “thought prompting” where the AI pauses for a moment before answering. Essentially the AI will “think” before speaking – just like a human does in milliseconds.

OpenAI is releasing new features fast and furious, Google is taking a more measured approach to their AI rollouts. They may appear to be behind but insiders are confident in Google’s technical abilities.

Google’s approach to AI is different from others. While some companies release products fast, Google prioritizes ethics and user trust. This cautious approach may seem slow but Google is keeping its reputation for responsible AI development. Experts acknowledge this methodical approach by saying Google has the technical chops to back it up and they are still in the AI race.

In July Google showed off two impressive models – AlphaProof and AlphaGeometry 2 – that can solve advanced math problems better than some of the best human minds. They got 4 out of 6 questions right in the International Mathematical Olympiad where students tackle tough topics like algebra and geometry. These results show Google’s focus on AI that can do advanced mathematical reasoning, a key factor according to DeepMind CEO Demis Hassabis.

And they showed off Astra, an AI assistant that can “see” through a phone’s camera and help users in real-time by answering visual questions – like helping someone find their lost glasses. This tech will be integrated into Google’s flagship AI model, Gemini, later this year.

Demis Hassabis, head of Google DeepMind, says advanced mathematical reasoning is key to truly intelligent AI. He says machines need to be able to solve complex problems with human-like reasoning to be intelligent.

OpenAI and Google are both pushing the limits of AI, the race is on to make machines think like humans. We don’t know when Google’s new model will be out but it’s clear they’re not going to let OpenAI get ahead. The battle is far from over.

Subscribe

Related articles

First IBM Quantum Data Center in Europe Opens; Will Include IBM’s Most Performant Quantum Systems

Alongside German Chancellor Olaf Scholz, senior European government officials and...

Pearson Unveils Generative AI Foundations Certification

AI proficiency is emerging as a must-have skill across...

Three-time Spacewalker Josh Cassada to Retire from NASA

NASA astronaut Josh Cassada retired on Oct. 1, after 11 years of...

Author

Tanya Roy
Tanya Roy
Tanya is a technology journalist with over three years of experience covering the latest trends and developments in the tech industry. She has a keen eye for spotting emerging technologies and a deep understanding of the business and cultural impact of technology. Share your article ideas and news story pitches at contact@alltechmagazine.com