Google DeepMind Will Release A Stronger Chatbot Than ChatGPT

JAKARTA - Google's research lab specialist in Artificial Intelligence (AI), DeepMind admitted that he was creating a more sophisticated competitor from ChatGPT.

Dubbed Gemini, chatbot uses techniques from AlphaGo, the first DeepMind AI system to beat a professional human in the Go board game in 2016, and leverage impressive capabilities of the Big Language Model (LLM).

In an interview with Wired, CEO and co-founder DeepMind Demis Hassabis said Gemini was designed to develop a chatbot that exceeds the current market leader's ability, OpenAI's ChatGPT, as well as comes more responsibly.

If all goes well, Hassabis said Gemini would have the ability to plan or solve problems and analyze text.

At a high level, you can view Gemini as a combination of some AlphaGo-type system power with the extraordinary language skills of a large model. We also have some new innovations that will be very interesting," said Hassabis.

In addition, Gemini will combine natural language processing and reinforcement learning. According to Wired, strengthening learning involves awarding AI systems for certain behaviors and or punishing unwanted behavior.

With the aim, teaching which behavioral systems will be shown in certain situations. Currently, it says Hassabis Gemini is under development, a process estimated to last several months and could cost tens or even hundreds of millions of dollars.

Launching TechCrunch, Tuesday, June 27, Gemini is not DeepMind's first breakthrough at LLM. Last year, the company introduced facter, a chatbot designed to provide an unsafe or inappropriate answer to the question. It is claimed, Didest's personal beta version will come this year.