AI can lie convincingly

A study by the University of Zurich shows artificial intelligence is capable of lying very persuasively – which could be a threat to democracy.

Researchers spoke to around 700 people to see if they could tell the difference between a tweet created by AI or a human.

Artificial tweets were found to be accurate and comprehensible. But the same technology can also be used to spread disinformation. 

Topics such as vaccine safety, covid, flat earth theory and climate change were used to test – and it was found AI had an ‘unsettling’ ability to spread false information highly effectively.

There are cross party talks on how best to manage the risks going on in Bern.

More from Bitesize News

  • Geneva could foot the G7 bill

    Geneva fears it will be left with the security bill from a G7 summit just across the border in Évian, according to 24 heures.

  • US NGO discovers Swiss child abuse cases

    Switzerland relies too heavily on an American NGO to detect child sexual abuse online, raising questions about the country's ability to protect children on its own.

  • French singer banned by Paleo

    The Paléo Festival in Nyon says it will no longer invite French singer Patrick Bruel, after a former volunteer accused him of inappropriate behaviour.

  • Cars going uninspected

    The cantons are struggling to keep up with mandatory vehicle inspections, leaving more than 550,000 checks overdue nationwide.

  • War may lead to rent rises

    Rents may rise again, says the bank Raiffeisen.

  • Not a heatwave, yet

    Summer has arrived, with temperatures climbing past 30 over the long weekend. Basel reached 31 and Sion hit 32.4, but MeteoSwiss, says this does not yet count as an official heatwave.

Download our app

  • Available on the App Store
  • Available on Google Play