news

한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina

Beijing, July 15 (Xinhua) -- If artificial intelligence can detect lies, would you use it to detect lies? A new study by an international team found that although using artificial intelligence to detect lies can help "separate the truth from the false", it also has certain negative effects and requires the establishment of a corresponding legal framework for supervision.

The University of Würzburg in Germany recently issued a press release saying that a team of researchers from the university developed a special artificial intelligence lie detection algorithm based on Google's open source large language model and related data sets and recruited human volunteers to participate in the test.

It is reported that in text-based lie detection, the accuracy of artificial intelligence is far higher than that of humans. Without the support of artificial intelligence, people would hesitate to accuse others of lying; with the support of artificial intelligence, people are more likely to express their suspicion of lies. About one-third of the subjects would seek help from artificial intelligence when identifying lies, and most of them would follow the algorithm's suggestions.

The researchers said that the study, published in the journal Cross Science, showed that while the use of artificial intelligence lie detection may encourage people to be more honest in their communications, it may also seriously undermine social harmony. After all, if people more frequently question each other's lies, it will increase distrust between people. Therefore, they called for the establishment of a comprehensive legal framework to make good use of new technologies while reducing the risks they may bring. (End)