Phones rang in the state of New Hampshire ahead of primary elections. Joe Biden’s voice was heard over the line. “We know the value of voting Democrats. It’s important that you save your vote for the November election,” the voice said.

“Voting this Tuesday only enables the Republicans in their quest to elect Donald Trump again. Your vote makes a difference in November, not this Tuesday.” Between 5,000 and 25,000 calls were made. The Biden administration never initiated any of those calls, and the voice was a deepfake — generated through artificial intelligence (AI) and mastered to sound like President Biden.

This example is part of a broader trend where AI-generated content was used to mislead voters during election periods. In Bangladesh, feeds of ‘international’ news channels were created using AI ahead of elections.

While imposing stringent laws or criminalising disinformation are pegged as solutions, such measures could inadvertently criminalise free speech and be used to suppress legitimate discourse. Therefore, the answer to combating disinformation isn’t as straightforward as ‘enacting a law against fake news’, or erecting a ‘national firewall’. It requires a careful balance between regulation and the preservation of free expression.

Read more here.