Google took its time to jump better
You know LaMDA or maybe you have already heard of it. Acronym for Language Model for Dialogue Applications, LaMDA is a language model like ChatGPT. He could notably talk about history, politics, invent, discuss like a human, do everything that the now famous artificial intelligence of OpenAI can do. LaMDA has been around for more than two years, but Google didn’t want to make a public demonstration of it. This decision by the administration resulted in the resignation of its designers: Daniel De Freitas and Noam Shazeer. They left the ship in 2021 to found their own company.
But then, why did the web giant not reveal its language model to take precedence over its competitors? What everyone could take today as a bad decision was nevertheless good in the eyes of Google and its chairman and CEO Sundar Pichai. Since the arrival of the CEO in 2015, he has put in place certain principles that guide the operation of the company so as not to reveal biased products.
To this effect, Google had judged that LaMDA “did not meet the ethical and security standards” of the company. Also, the product was not ready for large-scale use. By refusing to launch the chatbot, Google did not want to wither further, he whose actions are scrutinized very closely by state bodies.
SEE ALSO: Artificial intelligence is as dangerous as the atomic bomb, according to the ex-Google boss
Avoiding the classic problem of language models
Faced with the eagerness of Microsoft, which is completely changing its plans and betting everything on AI and OpenAI, whose product is gaining ground, Google did not hesitate to present its AI Bard in February 2023. Nevertheless, the company has still refused to open it for consumer use at this time. The Mountain View giant continues to work on its technology. The goal is to launch a fully ready product that can overtake the competition, as has been the case with its search engine.
Ending word
According to the statements, Google does not want to launch a product that would take over common problems in AI that are currently in the market. Indeed, ChatGPT and its friends tend to invent answers to questions they don’t understand. Google’s solution would be to allow its AI to recognize its ignorance in such cases, rather than mislead the user.
