Do you have a headache or is it a sinus infection? What does a stress fracture look like? Do you have to worry about pain in your chest? If you are looking for these questions about Google, the answers can be written by artificial intelligence.
In May, Google deployed a new feature called IA preview Who uses a generative AI, a type of automatic learning technology trained on information on the Internet and produces conversational answers to certain research questions in a few seconds.
During the weeks that followed the launch of the tool, users encountered a wide range of inaccuracies and strange responses on a range of subjects. The company later seemed to go back The functionality of certain research because it has tried to minimize these errors.
Regarding answers to health issues, the experts said that the issues were particularly high. Technology could indicate people to healthier or medical care habits, but it also has the potential to provide inaccurate information. AI can sometimes make facts. And if his answers are shaped by websites that are not based on science, he could offer advice that goes against medical advice or have a person’s health risk.
It has already been shown that the system produced bad answers Apparently based on defective sources. When asked “How many rocks I have to eat”, for example, IA glimps have told certain users to eat at least one rock per day for vitamins and minerals. (The council was scratched from Oniona satirical site.)
“You cannot trust everything you read,” said Dr. Karandeep Singh, director of UC San Diego Health. In health, he said, the source of your information is essential.