According to DeepMind, Google’s artificial intelligence (AI) subsidiary, the company’s future AI system will “be more capable” than OpenAI’s ChatGPT.
Gemini AI uses similar textual processes to ChatGPT
Demis Hassabis, the CEO of DeepMind, according to a report of Wired stated that the company’s Gemini AI system will be an advanced language model (LLM) that uses similar textual processes to ChatGPT. However, it will be enriched with new functionalities such as scheduling and troubleshooting.
The development of the system will build on the technology and techniques applied in AlphaGo, an earlier AI creation by DeepMind from 2016.
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.
In addition, he indicated that there will be “exciting new innovations” that he is excited about. Gemini was introduced during the Google I/O developer conference in May, where it was presented alongside other AI products.
Launch still several months away
Hassabis stressed that Gemini’s official launch is still several months away and the development process will ultimately require tens to hundreds of millions of dollars.
According to him, he sees no reason why the functionalities they develop with Gemini should not work.
In addition, DeepMind’s CEO stressed the “urgent” need for further research into the potential risks of more advanced AI systems. Earlier, Hassabis had already indicated that human-level AI could possibly appear before 2033.
The news about Gemini comes shortly after the launch of Google’s own AI chatbot, Bard, which competes with OpenAI’s much-discussed ChatGPT. However, Bard has faced internal misgivings from former Google employees regarding the “societal concerns”.
The launch of Google Bard in the European Union has already been blocked because regulators in Ireland claim the company has not submitted the required documentation.