Saturday, June 1, 2024

Criticism of Google's AI Overviews for Spreading Misinformation

Gemini large language models complement Google's AI Overviews, which have drawn criticism for producing erroneous and perhaps dangerous search results. In response to the criticism on query interpretation and depending too much on dubious sources like Reddit, Google has confessed mistakes. Reducing satire and user-generated material, the corporation has tightened policies for important subjects such news and health, thereby improving triggers for artificial intelligence overviews. Deactivating the function allows users to reduce the possibility of false information. Critics warn about counting only on artificial intelligence for search results even with best attempts to increase accuracy. Google aims to be able to reach more than a billion people by year's end and is considering running "Sponsored" commercials to profit from it. Understanding the requirement of responsible AI deployment, the company has revamped its solutions to provide accuracy, reliability, and openness first priority. Google has kept its AI overviews tool while noting in certain answers incorrect information and harmful material like unsafe advice and conspiracy theories. Among the guidelines used to avoid such errors are restrictions on user-generated material and improvement of detecting mechanisms for meaningless searches. Stressing the dangers of prejudice, false claims, and disinformation, AI specialists have recommended against depending only on responses produced by artificial intelligence.


Summary News

Gemini big language models provide Google's AI Overviews, which have been criticized for producing inaccurate and maybe dangerous search results. The corporation has come clean about mistakes in query interpretation and over relying on dubious sources like Reddit. Stricher rules for important subjects, declining of satire and user-generated material, and improved triggers for AI overviews have been developed to help with this. Users may now disable the functionality in order to lower false information risk. Google wants to profit on the feature available to around a billion people by year's end with "sponsored" advertising.


Source: Google’s AI Recommended Adding Glue To Pizza And Other Misinformation—What Caused The Viral Blunders?
Source: Google Rolls Back A.I. Search Feature After Flubs and Flaws
Sources: Forbes, The New York Times

No comments:

Post a Comment