Is AI-generated content too believable?

51401

Total votes

14077

5-11 votes

37324

11-16+ votes

Last week, voters were exploring the world of AI-generated images, audio, and video in celebration of British Science Week. They considered the increased sophistication of content created using artificial intelligence and how this is influencing everything from popular culture to politics.

16+ & College voters were asked: “Is AI-generated content too convincing?”, while Secondary voters explored: “Is AI-generated content too believable?”, Primary 7-11 voters discussed: “Are AI-generated images too believable?” and Primary 5-7 voters considered: “Can you tell when photos are real?”

Age 5-11 voter

AI is getting more and more sophisticated and so unreal things would be harder to spot. Nobody would think something is unreal unless they were told that it was and they start scrutinizing it themselves.

Age 11-16 voter

It may not be too believable now, but it will continue to improve. It can become very dangerous and can lead to a massive distrust of any kind of material, whether it is real or not.

Age 16+ voter

AI should have necessary filters to protect human produced content including: creative arts industries, writing and game development.

50% of 5-7-year-olds argued that they can tell when photos are real.

Our youngest voters were discussing the question, "Can you tell when photos are real?"
The result was mixed, with 50% voting 'Yes' and 50% voting 'No'.

56% of 7-11-year-olds said that AI-generated images are too believable.

7-11-year-olds were debating the question, "Are AI-generated images too believable?"
Just over half, at 56% voted 'Yes'.

54% of 11-16+ year old students agreed, saying that AI-generated content is too convincing.

Secondary, College and 16+ students were exploring the question, "Is AI-generated content too believable(11-16)/convincing(16+)?"
Just over half, at 54% voted 'Yes'.

Some students argued that AI images will inevitably become more believable as the technology advances.

Many 'No' voters argued that, whilst they can often tell when an image is AI, it is likely that this will be more difficult as time goes on.