Google DeepMind is apparently building a ChatGPT killer AI bot

When it comes to artificial intelligence (AI), OpenAI’s ChatGPT has created a buzz around the world, and as per what Sam Altman, the creator of ChatGPT, claims, it took around $100 million to create the AI bot. To rival such an innovation seems like a challenge in itself, but Google is here with their innovation. Google’s research lab, DeepMind, has asserted that their next large language model will rival and maybe even turn out to be better than OpenAI’s bot.

A champion Go player was defeated by an artificial intelligence programme named AlphaGo from Google’s DeepMind AI lab in 2016, making history. Now, DeepMind cofounder and CEO Demis Hassabis claims that his company’s developers are employing AlphaGo-related methods to create Gemini, an AI system that will be superior to OpenAI’s ChatGPT.

Gemini, a large language model from DeepMind that uses text and is currently under development, is comparable to GPT-4, the engine that drives ChatGPT. But according to Hassabis, his team will integrate that technology with AlphaGo-specific strategies in an effort to offer the system new talents like planning or problem-solving skills.

Demis Hassabis expressed, “At a high level, you can think of Gemini as combining some of the strengths of AlphaGo-type systems with the amazing language capabilities of the large models.”

Hassabis further added, “We also have some new innovations that are going to be pretty interesting.” At Google’s developer conference last month, when the company unveiled a number of new AI projects, Gemini was first alluded to.

AlphaGo is built on a method DeepMind invented called reinforcement learning, in which computer programmes are taught to tackle challenging tasks requiring decision-making, such as in video games or the game of Go, through repeated efforts and evaluation of their performance. Additionally, it explored and stored potential moves on the board using a technique called tree search.

It’s possible that language models will make their next major advancement by using computers and the internet more frequently. Gemini is still being developed, which will take several months, according to Hassabis. Tens or even hundreds of millions of dollars may be required.

It is being asserted that once Gemini is completed, it could pose as a tough rival to OpenAI’s ChatGPT and other generative AI technologies. Google has introduced its AI tools and services in the space, but it seems a bit rushed since all of it came in the face of ChatGPT’s launch. Google has not only introduced its generative AI, Bard, but also put AI in its search engine and many of its services.

To level things up a notch, Google combined Hassabis’ unit DeepMind,with its primary AI lab, Brain, to create Google DeepMind. According to Hassabis, the team is working to combine two powerhouses that have served as the foundation for recent AI advancements. Hassabis expressed, “If you look at where we are in AI, I would argue that 80 or 90 percent of the innovations come from one or the other. There are brilliant things that have been done by both organisations over the last decade.”

Reports suggest that Hassabis has quite a bit of experience organising and directing AI gold rushes that have the potential to baffle the tech giants. However, last time, Hassabis may have sparked a frenzy around him.