Google returns AI search answers after asking users to eat gum

SAN FRANCISCO — Google said it is reducing the use of AI-generated answers in some search results after the technology made high-profile errors, including asking users to put glue on their pizza and saying Barack Obama was Muslim.

Google began placing artificial intelligence answers at the top of search results for users in the US two weeks ago, but users and search engine experts on Thursday began noticing that far fewer queries were triggering an AI response compared to previous days.

Liz Reed, Google’s head of search, confirmed Blog On Thursday afternoon the company scaled back some of the AI ​​responses, which it called “AI overviews.” The company has reduced the use of social media posts as raw material for AI answers, suspended some answers on health-related topics and added “trigger restrictions to questions that demonstrate AI overviews are not useful,” Reid said in the filing. . The company has made more than a dozen technical changes aimed at improving the system, Reed said.

The change is the latest example of Google launching an AI product with fanfare, then retracting it after it goes wrong. In February, the company blocked users from creating images with its AI imaging tool after conservative commentators accused it of anti-white bias.

The tech industry is in the midst of an AI revolution, with start-ups and big tech giants trying to find new ways to monetize the technology by incorporating it into their products. Many tools are launched before they are ready for prime time, as companies seek to be the first to market their products and present themselves as cutting-edge.

See also  President Biden calls China's Xi a dictator

get caught

Short stories for quick information

Google, which invented most of the technology-based breakthrough AI tools like ChatGPT, is trying to prove to investors, consumers and its own employees that it can still play an important role in the industry. At its I/O conference this month, the company made more than 100 AI-related announcements.

The big one is that it confirmed that it will start publishing AI-generated answers in search results for most of its users. Google has been testing AI answers for a year with a select group of users, but as it adds them to more search results, most people will now start interacting directly with AI on a tool they use every day.

The technology works by reading websites displayed in Google search results and summarizing them into multi-paragraph answers. Publishers are accused of harming their businesses by taking their content and retargeting it directly to users in search results, potentially losing valuable web traffic.

But journalists, search engine experts and social media users quickly began to spot problems with the answers. Some of the answers were funny and others were concerning. They were also featured in important questions including health related questions.

One answer corrected by Google said that you need to drink a lot of urine to pass a kidney stone. John F. Kennedy graduated from the University of Wisconsin at Madison in six different years, three of which were posthumous.

Google tried to test the tool as much as possible before the wider release, but Reid said the full-scale rollout exposed many scenarios the company hadn’t prepared for.

See also  Brazil lifts ban on Elon Musk's X after paying $5 million fine

“There’s nothing like millions of people using this feature,” Reid said.

Leave a Reply

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