Last weekend, the Internet was obsessed with a picture of Pope FrancisThe head of the Catholic Church since 2013. An Argentinian, he is the first non-European to hold the position since the 8th Century.. It showed him wearing a white puffer jacket. Commentators raved about his chic get-up.

All was not as it seems, however. The "Balenciaga Pope"¹ was a deepfake: a false image generated by AIA computer programme that has been designed to think. . We had all been taken for fools. 

Many think this is only the beginning. On Wednesday, over 1,000 tech industry figures - including Twitter owner Elon MuskA South African-born entrepreneur whose companies have included the online payment service PayPal. and Apple founder Steve Wozniak - signed an open letter calling for all AI labs to pause their research.

The letter stated: "Recent months have seen AI labs locked in an out-of-control race to develop and deploy ever more powerful digital minds that no one - not even their creators - can understand, predict or reliably control".

There are many reasons to be alarmed. One is AI's ability to spread misinformationIncorrect or misleading information unintentionally presented as fact. It can be contrasted with disinformation or deliberate lies.. Chatbots could spread conspiracy theories. Election campaigners could fill the internet with deepfakes of opponents. 

Last week, a right-wing influencer created a fake video of Joe BidenThe president of the USA, elected in 2020. ordering US soldiers to Ukraine. He said it was a fake. But many who shared it did not. The same happened with an AI-generated image of Donald TrumpAn American television personality who became US president from 2016 to 2020. He is running to be president again in 2024. being arrested.

Imagine this power in the hands of an authoritarianEnforcing strict obedience to authority. government. According to misinformation expert Gordon Crovitz: "Crafting a new false narrative can now be done at dramatic scale, and much more frequently."

This process might have already begun. Last year, a video emerged in which a deepfake of UkrainianUkraine is an eastern European country that borders Russia. The country was part of the Soviet Union until it became independent in 1991. In recent years it has come into conflict with Russia, which seized parts of its territory in 2014. president Volodymyr Zelensky asked his army to surrender. It was quickly revealed as fake. But the next one might not be.

Others believe we worry too much. AI software can be programmed to defend against fakery. ChatGPT-3 went crazy and wrote offensive messages; ChatGPT-4 has been coded to avoid these outcomes. Developers can limit how AI can be used. Governments can restrict them.

AI-generated text and images are far from perfect. Deepfake videos do not quite move in the same way as people. The Balenciaga Pope's crucifixA cross with an image or representation of Jesus on it. hung by only one side of a chain, his eyelid melted into his glasses, and his coffee cup was held wrong. Anyone who looked at the picture closely would see these mistakes.

Very: Social media was created to share photographs with friends. It led to cyberbullying and conspiracy theories. AI is much more powerful. There is every reason to think the effects will be even worse.

Not much: We were fooled by the Balenciaga Pope. But when it comes to serious matters like political campaigns, people are much more careful. As long as we keep our wits about us, we will be fine.

Or... It is too late. AI has already become incredibly powerful. No-one, from governments to tech companies, has lifted a finger to stop it. Rather than panic, we need to accept this new reality and adapt.