Word of the day
Google has launched a new artificial intelligence (AI) model that it claims will "answer questions more carefully" because its "functionalities are more advanced".
AI content generators sometimes invent things that developers call hallucinations.
The new Gemini model has been tested to solve problems in 57 areas, including mathematics and the humanities. "This is the new era of artificial intelligence," said Sundar Pichai, head of Google.
Earlier this year, Google launched its own version of the chatbot – Bard – calling it an "experiment".
During the presentation, the program made an error when answering the space question.
Google claims that the new model is "the most capable" and will outperform humans in solving intelligence tests.
Gemini can recognize image, text and sound, but it is not a product in itself.
It is a program that will be integrated into existing Google tools, including search, but also into the Bard chatbot.
Gemini appears to have set the standard, showing that the program can read sources other than text content, such as images, according to Chirag Dikeit, an analyst at Gartner.
This could, he adds, "affect the development of innovations that would likely transform artificial intelligence."
So far, Google has struggled to attract users like OpenAI is doing with ChatGPT.
The company claims that the new version is better than the OpenAI platform GPT-4 that runs ChatGPT in terms of performance.
A more powerful version of the OpenAI software is expected to appear in 2024.
Sam Altman, CEO of ChatGPT, says the new products will make the current ones look "like distant cousins".
It is not known whether the recent turmoil in the company, due to which Altman was fired and returned to his job after a few days, will affect the launch of the new version.
The company also has a new competitor, Elon Musk's hAI, owner of social network X, which is trying to raise up to one billion euros to invest in research and development.
And the Chinese company Baidu has joined the race in the field of artificial intelligence, developing its own artificial intelligence products.
But as technology advances, so do fears of its potential harm.
Governments around the world are trying to establish rules or even laws to prevent possible future risks from artificial intelligence.
This topic was discussed at the summit in Great Britain, held in November.
A declaration was signed calling for the safe development of artificial intelligence.
And Britain's King Charles III pointed to the potential dangers that need to be dealt with "urgently, unitedly and jointly".
Comments
Post a Comment