Google executive warns of ‘hallucinating’ chatbots

The fire for AI tools is growing and by all indications it shouldn’t end anytime soon. One of the biggest culprits for this is undoubtedly ChatGPT, an artificial intelligence chatbot launched by the OpenAI company in November 2022. The platform gathered millions of users in a very short time and became a digital phenomenon. But, According to Google, people should make it easy when it comes to chatbots.

In an article published in a German newspaper called Welt am SonntagPrabhakar Raghavan, senior vice president of Google and head of Google Search (the company’s search tool), warned last Saturday (11) of the downsides of artificial intelligence in chatbots, stressing that sources can often be misleading.

“This kind of AI that we’re talking about right now can sometimes lead to something we call hallucinations. […]. This is expressed in such a way that the machine gives convincing answers, but was completely invented,” explained the manager. In this sense, according to Raghavan, the challenge will be to prevent or reduce this effect as much as possible.

Google chases the competition

Since the launch of ChatGPT, Google has struggled to stay behind in the field of artificial intelligence. The company has yet to release an AI product that rivals OpenAI’s chatbot, which has plagued company executives, although it has presented many tech-related studies over time.

Alphabet Inc., the holding company that includes Google, has even unveiled its own chatbot called Bard, which it announced on Feb. However, the tool provided false information in a promotional video and created a major blunder that cost the company an estimated $100 billion market capitalization.

Bard continues to be tested, but there are no updates yet or an official release date for the general public. “Obviously we’re in a hurry, but we also have a great responsibility. We certainly don’t want to mislead the public,” Raghavan concluded.