sourcegraph
March 29, 2024

Industry experts believe that as technology improves, Google must decide whether to overhaul its search engine and make full-fledged chatbots the face of its flagship service.

Google has been reluctant to share its technology widely because, like ChatGPT and similar systems, it generates false, toxic and biased information. LaMDA is only available to a limited number of people through the experimental app AI Test Kitchen.

Google sees it as a struggle to deploy its advanced artificial intelligence without harming users or society, according to a memo seen by The Times. At a recent meeting, a manager acknowledged that smaller companies were less concerned about releasing the tools, but said Google must step in to compete or the industry could survive without it, according to meeting recordings obtained by The New York Times. continue to move forward.

Other companies have similar problems. Five years ago, Microsoft released a chatbot called Tay that spewed racist, xenophobic and other foul language and was forced to immediately remove it from the internet — never to be seen again. In recent weeks, Meta shut down an updated chatbot for many of the same reasons.

Google intends to release the technology that powers its chatbots as a cloud-computing service for external businesses and may integrate the technology into simple customer support tasks, executives said in a recorded meeting. It will maintain trust and safety standards for official products, but will also release prototypes that do not meet those standards.

It may limit these prototypes to 500,000 users, warning them that the technology may make false or objectionable statements. ChatGPT (which produces a similarly toxic substance) has been used by more than a million people since its launch on the last day of November.

“A cool demonstration of a dialog system that people can interact with in a few turns and feel exciting? It’s a good step, but it’s not something that’s really going to change society,” said Google Brain, an artificial intelligence lab. Zoubin Ghahramani, head of ChatGPT, said in an interview with The Times last month ahead of ChatGPT’s launch. “It’s not something that people can reliably use every day.”



Source link

Leave a Reply

Your email address will not be published. Required fields are marked *