Google presents new AI model that ‘beats GPT in almost every test’

Google has unveiled a new AI model it claims beats rivals like ChatGPT at most tasks following the company’s “biggest science and engineering project ever”.

Gemini is described as multimodal, meaning it can understand text, audio, video, images and computer code.

It is built by cross-functional teams Googleincluding London-based AI research unit DeepMind.

DeepMind’s CEO, Demis Hassabis, hailed it as “a significant milestone in the development of artificial intelligence”.

Gemini has three levels of capability, and the company said the highest ultra version, designed to perform “very complex tasks,” outperforms similar OpenAI’s GPT models in 30 out of 32 academic benchmarks widely used in AI research and – development.

A promotional video showed it assessing and providing feedback on a student’s physics homework.

It will not be released to the public until early next year.

The second level is pro, which is already rolled into Google’s Bard chatbot. It is being launched in more than 170 countries including the US, but has yet to be approved by regulators in the UK and the EU.

Google said the upgrade will make Bard “more adept at things like understanding and summarizing, reasoning, brainstorming, writing and planning”.

Gemini’s third tier is nano, which is designed to run natively on devices like Google’s Pixel smartphones. It will do things like summarize voice recordings and suggest replies in messaging apps.

Read more:
How chaos unfolded at ChatGPT’s OpenAI
A year of chatbots that changed the world
‘Godfather of AI’ – and why is he so worried about his life’s work

The unveiling of Gemini is 2023’s latest major development in artificial intelligence.

It comes 10 months after the release of Bardwhich was widely reported to have been fast-tracked after the search giant was alarmed by the success of OpenAI’s ChatGPT in November 2022.

Wall Street welcomed the company’s bullish announcement about Gemini, with shares in parent Alphabet rising 5%.

It is also the first major release of a new AI model since Britain hosted a world-first security summit last month, when Prime Minister Rishi Sunak announced that tech companies had agreed to share data from their internal tests with the UK government.

They are also required to share test data with the US government after President Joe Biden announced sweeping regulations designed to address concerns about AI.

The EU is expected to unveil its own legislative proposals soon, with negotiations this week.

According to the Associated Press, representatives from the bloc’s 27 member states and officials from the European Commission argued over the details of the AI ​​law for 22 straight hours before adjourning on Thursday.

Sky News has contacted Google for comment on its plans to share Gemini test data and the reasons why Bard’s upgrade has been delayed in the UK and EU.


Leave a Reply

Your email address will not be published. Required fields are marked *