Obviously, it’s worrying.
But there must be something wrong with me: I find it fascinating.
Not because I’d be delighted to see lying become industrialised, but because this technological revolution is forcing us, rather abruptly, to ask ourselves a question we may have stopped asking: why do we believe what we believe?
For over a century, photography and then video have enjoyed a special status. A photograph could lie, of course. It could be cropped, staged, taken out of context or retouched. But behind it all lay that extremely powerful idea: ‘it happened’.
Roland Barthes spoke of photography’s ‘ça-a-été’. Something had necessarily been in front of the lens.
Artificial intelligence is in the process of shattering this contract.
From now on, ‘I saw it on video’ will no longer necessarily mean very much.
And that is significant.
We are going to have to learn to look at an image as we should always have done: by asking ourselves who produced it, why it exists, where it comes from, in what context it is presented to us, and who stands to gain from us believing it.
In other words, AI may not just be creating a crisis of truth. It is also revealing just how badly we’d got into the habit of relying on it.
I tried, in my own very small way, to convey something along these lines to my daughters when they were children.
I used irony a great deal with them. I would tell them something with the utmost seriousness before letting them realise that I’d just been pulling their leg. It was playful, sometimes absurd, and behind the game lay a fairly simple idea: don’t believe me simply because I’m your father.
Check it out.
Think about it.
Ask yourselves whether what I’m saying makes sense.
And above all, allow yourselves to think that the adult standing in front of you might be wrong, joking, exaggerating or talking nonsense.
Of course, I didn’t want to teach them that nothing was true. That would be the exact opposite of critical thinking. I wanted to teach them that between believing everything and believing nothing, there is a much more interesting space: that of judgement.
Today, they are two intelligent young women, capable of forming their own view of the world, and I think that this little family-run ‘school of doubt’ may well have contributed to that.
That is precisely what the coming generations are going to need.
For the real danger of artificial intelligence may not be that we believe all the fake images.
The danger could be exactly the opposite: that we end up no longer believing any images at all.
Imagine that world for a moment.
A video shows police brutality? ‘It’s AI.’
A politician is filmed making scandalous remarks? ‘Deepfake.’
A photograph documents a war crime? “Computer-generated.”
A public figure is caught red-handed? “Fake.”
Once everything can be fabricated, anything that is authentic can also be presented as fabricated.
This is the fascinating and terrifying paradox of this revolution.
The fake becomes more credible at the very moment when the real becomes easier to deny.
We are therefore entering not so much an age of lies as an age of uncertainty.
And that will require a proper education.
There is much talk of teaching children how to use artificial intelligence. Very well. But above all, we need to teach them how to live in a world where it exists.
This means learning to trace a source, to compare different pieces of information, to identify the creator of an image, to understand the difference between evidence and impression, to recognise one’s own biases and, above all, to accept that phrase which has become almost unbearable in the age of social media: ‘I don’t know.’
We’ve got into the habit of forming an opinion on everything within thirty seconds.
Paradoxically, AI might force us to rediscover the patience of doubt.
And doubt is not a weakness of intelligence.
It is probably one of its prerequisites.
What worries me more than the advances in artificial intelligence is the prospect of a society made up of individuals incapable of questioning what appears on their screens. Because even before Midjourney, ChatGPT, deepfakes and video generators, propaganda already existed. Manipulation already existed. Faked photographs already existed. Rumours already existed.
Human beings did not wait for artificial intelligence to start lying.
They have simply given themselves extraordinarily powerful tools to do so.
So yes, the world that lies ahead is mad. Yes, it is unsettling. Yes, there will be political manipulation, scams, reputations destroyed by fake videos and probably crises triggered by images that never actually existed.
It would be naïve to deny it.
But I also reject this very human tendency to greet every technological revolution by immediately predicting the collapse of civilisation.
AI presents us with considerable problems. It also offers us a rather extraordinary opportunity: that of putting critical thinking back at the heart of our relationship with the world.
A photograph is not the truth.
A video is not the truth.
A journalist is not the truth.
An influencer is not the truth.
Artificial intelligence is obviously not the truth.
And neither am I.
These are sources, perspectives, documents, narratives and tools that must be tested against reality.
We are going to have to relearn something we perhaps should never have forgotten: seeing is not knowing.
Personally, this prospect worries me as much as it excites me.
Because a world in which we must question more closely what we are shown is certainly more uncomfortable.
But it can also become a wiser place.
Provided we do not replace credulity with paranoia.
Doubting everything does not mean believing that everything is false.
It means being willing to seek the truth.
And perhaps, in the world to come, the truth will no longer be what we see.
It will be what we take the time to verify.
