My phone rang at eleven o'clock one night. It was my sister's voice, and she was crying.
She said she was in trouble. She needed money, that night, immediately. And she asked me not to tell our mother.
For a moment I believed her completely. I was already on my feet, reaching for my wallet. Her voice has had that effect on me my whole life.
Then a second thought arrived — the thought that occurs to all of us now. It said: that might not be her. It might be a few seconds of her laughter, taken from a birthday video and fed to a machine. This happens. Most of us have read about this exact call happening to someone else.
So I did what you are advised to do. I asked her something only she would know — something from our childhood, written down nowhere. And she answered it correctly.
Here is the part I cannot stop thinking about. It did not help. I sat there with the wallet in my hand and realised that a good enough fake would know the answer too. Anyone able to fabricate her crying voice could also answer my question. There is no question I could ask that a good enough forgery could not pass. I understood that, and I sent the money anyway — not because I had proved it was her, but because I had run out of ways to be sure and decided regardless.
It was her, as it turned out. She was fine. The next day she was mortified that I had nearly not believed her.
But it did not end with the phone call. The following morning I found myself doubting a news clip — genuine footage of a real event — for no reason except that everything now looks equally uncertain. A year earlier I would have believed it without a second thought. This time I did not, just in case.
This is Philosophy for Us — philosophy for everyone, no degree required. This is the first of eight episodes on what may be the strangest problem of the present moment: how you know anything you did not witness yourself, now that the things we used to rely on to show us — a photograph, a recording, a voice on the phone — can be made to deceive us.
This programme has spent whole series close to home: what we owe a stranger; grief; the mind, and the machine that answers back. Tonight we begin with something that feels like a problem of our moment and no other — deepfakes, "do your own research," the whole anxiety about a post-truth world. It deserves to be taken seriously, which means declining to leave it as an anxiety. Because underneath this very current question there is a much older one, and it has been there the whole time.
Let me say plainly what I will not do. I will not give you a way to detect fakes. There is no test at the end of tonight — no tell in the audio, no checklist, no method that makes you certain. If that is what you came for, I would rather say so now than pretend otherwise at the end. Nor am I standing outside the problem: that was my wallet, my sister, my own shaking hands. What I can do is show you something about the question itself — that you have been answering a different question from the one you are stuck on, and answering it on an assumption you have never actually said out loud. Tonight we find that assumption. First, back to the phone.
Go back to the phone for a moment, because there is something in that story most of the worry about all this passes over.
When people warn you about fakes, they warn about one thing: that you will be deceived. Someone will clone a voice, fabricate a video, forge a document; you will believe it; and you will lose your money, or your vote, or your bearings. That is a real danger. It does happen. But it is the smaller half of the problem. The larger half was in the room with me that night.
The larger half is this. Once you know that fakes exist, you do not merely disbelieve the fakes. You begin to disbelieve true things.
Consider how it works. The existence of forgeries teaches a lesson: anything could be fake. That lesson is correct. But it does not stay in that measured form. Within a week it has broadened into "probably is." A clip reaches you, and a year ago you would have taken it as a plain fact about the world; now you think, who knows — it could be generated — and you decline to believe it. Not because you have any reason to think this particular clip is false; you have none. But because doubting costs you nothing, and believing might make you the fool. So you doubt. You doubt genuine footage. You doubt an honest source. You doubt something that actually happened.
That is the trap, and it is worth stating slowly, because it is the reverse of how it feels. It feels like becoming sharper — more skeptical, harder to deceive. What is actually happening is that you are losing things you already had: the true ones. The forgery did not merely let one false belief through. It trained you to reject true beliefs as well.
There is a name for what happens next, and it is older than the technology. Once people know that lies of this kind are possible, a liar gains an advantage he did not have before: he can dismiss a true statement about himself as just another fake. The genuine recording of what he did? It could have been doctored. The real photograph? Probably generated. He does not have to prove the true thing false. He need only point out that, these days, no one can tell. The mere existence of convincing fakes gives every liar a way to evade every true accusation. So fakes do not only plant falsehoods; they also protect the people the truth would otherwise have caught.
The damage runs in two directions at once, and the second is the deeper. A fake you believe costs you a false belief — bad, but limited: one wrong item in your head. A fake you have only heard about costs you something without limit. It places under suspicion every belief you hold on the strength of a recording, a report, or a source — not only the false ones but the true ones with them. You came to the problem of fakes worried that you would accept something false. The worse consequence is that you can no longer trust what was accurate all along.
And here is what makes it acute. The remedy people reach for makes it worse. Be more skeptical, they say — doubt more, trust less, assume nothing until it is proven. But this second, deeper damage is made of skepticism. The liar's advantage works only because you have learned to doubt. The true accusation dies only because you have been trained to say "could be fake" about everything. So the standard advice — trust less — is not the remedy. It is the very thing that produces it. The more thoroughly you absorb the lesson that anything could be fake, the more true things you lose, and the safer every liar becomes. Doubt was meant to protect you. It turns out doubt is also what is being used against you.
I am not saying you should believe everything. I am not saying that — and by the end of these eight episodes I will argue that believing everything is its own kind of disaster, with its own heavy cost. What I am saying tonight is narrower and stranger: that the obvious response to a world full of fakes — verify more, trust less, become harder to deceive — does not get you out. It is part of the problem itself. The person who ends up knowing the least is not the credulous one who believes the fakes. It is the one who, having been deceived once, has learned to believe nothing, and now lives among a great many true things unable to count any of them as true.
That was my position at midnight. I had not been deceived. My sister was real, the call was real, the question was answered. I had done everything correctly. And I was worse off than if I had simply believed her — because I had learned, that night, that there was no longer anything I could do to be certain; and the lesson did not remain with the phone call. It extended to the news the next morning. It will extend everywhere, if you let it, into everything you know secondhand.
Before we look for a way out, that has to be fully in view. The problem is not only that you might be deceived. The problem is that the correction you would reach for — doubt harder — removes the true things along with the false ones without your noticing, and gives liars protection while it does so. That is the shape of the difficulty. Now let us look at what people turn to in response, because nearly everyone turns to the same thing, and it is worth seeing clearly before we ask whether it works.
So what do you do? When you can no longer be sure, and yet you still have to know things — you cannot opt out of knowing; you have a life to run — what do you turn to?
I will tell you what I turned to, because I think it is what nearly everyone turns to, and it is worth naming plainly, since we will be living with it for some time.
I turned to checking. Verify it. Ask the question only she would know. Find a second source. Look for the original. Run the photograph through a reverse image search. Read past the headline to the study, and past the study to the data. Do not take it on faith — confirm it. Beneath all those moves is a single instinct: do not believe a thing until you have checked it for yourself. Believing without checking is how you are deceived; checking is how you stay safe. The fakes have made checking harder, so the conclusion seems obvious — check harder.
You have heard the popular version of this: do your own research. It is everywhere now, and it has become a kind of moral instruction — something you owe, a mark that you are no one's dupe. Do not simply accept what you are told; go and find out for yourself. Let me be fair to it, more fair than it usually gets, because there is something genuinely right in it. There is a real refusal in "do your own research" — a refusal to defer blindly, to believe a thing merely because a confident voice asserted it. That refusal is not foolish. Some of the worst things people have ever believed, they believed because an authority said so and they did not check. The impulse not to accept whatever one is told is a healthy one. Keep that refusal in mind; we will return to it.
But notice what the whole family of responses has in common, because this is what tonight is really about. Every one of them — verify, confirm, source it, do your own research — treats the problem as a problem of checking. As though the situation were this: you are a basically competent knower, on secure ground, and what the fakes have done is damage your ability to check. If that is the situation, the repair is obvious: restore the ability to check, sharpen it, be more careful — recover the capacity that was under attack, and you are back on secure ground.
That account feels like rigour. It feels like the responsible, hard-headed response — the opposite of being a dupe. I do not want to take that feeling away from you yet, because there is a real position here, a serious one, with serious people behind it, and next time I will put it to you at full strength. There is a whole tradition that says this: that a wise person believes things in proportion to the evidence, treats no one's word as worth anything without independent grounds for it, and regards an extraordinary claim as requiring extraordinary proof. On that view the deepfake era is not a catastrophe at all. It is a vindication — the world finally confirming what careful people have always said: that you should never have been taking things on trust in the first place, and now you have no choice but to do the verifying you always ought to have done. I told you, says the careful person. Check everything.
I am not going to argue with that tonight. I will let it stand, because it deserves to stand, and because if I knock it down too quickly you will think I am being clever rather than honest. So let it stand. The verify-everything answer stands, in its strongest form, and there is a real philosopher behind it whom we will meet next time.
I will only raise one question about it, and then leave it alone. The question is this: can you actually do it? Can anyone? Not should you — can you. Can you live a single day by the rule "believe nothing you have not checked for yourself"? I tried, for about a day, after the phone call. I tried to notice every time I was taking something on someone's word, and to mark it for checking. I will tell you what happened, but not yet — because what happened is the thing this whole series is built to show, and it would be cheating to hand it over tonight. For now, keep the question in front of you. The answer that feels like rigour — check everything, trust nothing until proven — sounds unanswerable. But there is a difference between a rule that is true and a rule you could actually live by, and I am not yet sure this one is both.
So this is where we are. You have a world in which doubting harder only spreads the damage. And you have one response — really the only reflex most of us have — which says the way out is to check more, verify more, do your own research, become harder to deceive. That response rests on a picture of who you are: a competent knower, on secure ground, whose ability to check has been damaged and now needs repair. Next time we will take that picture and put the strongest possible philosopher behind it, and let him argue that the picture is not merely comforting but correct — that you were always meant to be the one who checks, and that your only mistake was becoming lazy about it.
But before we turn to him, one more thing has to be in plain view. That picture — the competent knower, the secure ground, the damaged ability to check — has an assumption buried inside it. An assumption so deep that you have never had to state it aloud, which is exactly why it can govern your whole response without your noticing. Let me bring it up and state it plainly. That is the last thing we do tonight.
Here is the assumption. I will state it as plainly as I can, and test it against your own response as I do, because the whole point is that it should feel obvious. If it feels obvious, that is the assumption at work.
The assumption is this. Real knowledge is what you have checked for yourself. That is the standard — what you have seen with your own eyes, worked out with your own reasoning, confirmed with your own hands. Everything else — everything you have merely been told, everything you hold on someone else's word — is a lesser grade of knowledge. It is provisional. You hold it provisionally, at your own risk, until the day you can verify it and promote it to the real thing. Real knowledge is what you establish for yourself; being told is the second-best you accept when establishing it yourself is not available.
Say it back to yourself and notice how little it seems to need defending. Of course the things I have checked are more securely mine than the things I have only been told. Of course taking someone's word is the second-best option, the thing you settle for. It is so obvious that it barely counts as a belief at all. It is less a belief than the basis on which your other beliefs rest.
And that is exactly why it has been governing your whole response to the fakes without ever appearing openly. Consider what the assumption does for you. If real knowing is checking for yourself, then the fakes are a pure catastrophe, because they wreck your ability to check, and checking was the whole enterprise. Your good knowledge, the firsthand kind, is under direct attack. So of course you turn to verifying harder; of course "do your own research" feels like the only adult response. The assumption tells you that you are someone who checks, dislodged from secure ground, and that the only way back is to check better. Every move we have discussed tonight follows from this one buried idea: that you are, fundamentally, a person who knows things by checking them, and that trust is the flimsy material you fall back on when you cannot.
I am not going to argue with that yet — that is next time, and the time after. Tonight I only want it stated and set out where you can see it, because in a moment I am going to ask you to keep hold of it, and you cannot keep hold of something you cannot see.
Before we close, let me show you the assumption at work in a small case that has nothing to do with your sister or a forged video, so that you can catch it yourself rather than take my word for it — which would be a little odd, given the subject.
Here is the case. You are at dinner, and someone states something you are unsure about. Violent crime has risen sharply. Or a certain medication is dangerous. Or the bridge project is three times over budget. And there it is — the reach. Your hand wants your phone. Look it up. Check it. Do not simply take their word, and do not simply trust your own hunch either; confirm it, find out for yourself, settle it with the facts. That impulse feels good. It feels like the responsible thing, the rigorous thing, the opposite of being the person who believes whatever is asserted most loudly. That feeling — that looking it up is what being a careful knower consists in — is the assumption itself. Checking is knowing; checking is the rigour; everything unchecked is suspect.
Now slow down, because there is one small observation to make here, and then I will leave it alone. You reach for the phone to look it up. What is on the phone? An article someone wrote. A study you did not run, conducted by people you have never met, summarising data you will never see, written up by a journalist you are also trusting, placed at the top of your screen by a company you are also trusting. When you "look it up" and "find out for yourself," what you are doing, every single time, is trading the word of the person at dinner for the word of some other people you cannot see. You did not step outside taking people's word. You changed whose word you were taking.
I am not going to make anything of that tonight. I am not going to tell you it means checking is pointless, because it does not, and I think checking matters a great deal — we will spend a whole hour on someone who has built a serious philosophy out of that. I am only asking you to notice the point. What felt like leaving trust behind and standing on your own — looking it up, doing your own research — was itself, when you watch it closely, another act of trust. Only a quicker one, directed elsewhere.
The question to carry out of tonight, unanswered, is this. If even your checking turns out to be a form of trusting — if "find out for yourself" is largely "take a different stranger's word for it" — then how much of what you know was ever really checked, in the firsthand, establish-it-yourself way the assumption calls the only real kind? How much of what you know is built that way? The assumption told you that you stand on ground you checked for yourself, and we have just caught it being inaccurate in a small case. The next question is how small the inaccuracy really is. And I do not think it is small at all.
Let me draw it together, and say plainly what tonight leaves unresolved.
The deepfake feels entirely new, and in one respect it is: no one could fabricate your sister's crying voice a few years ago, and now they can. But notice what it points to. It is not, at bottom, a problem with cameras. The camera was only ever showing you a small part of what you claim to know. The real question underneath the whole anxiety — the one no one stops to ask — is this: how much of what is in your head did you ever take in through your own eyes — and how much of it arrived the way the phone call did, secondhand, on someone's word? We are going to return and do this properly, with your own beliefs set out in front of you, and I think the count will surprise you. Tonight it is enough that the question is in view.
For now, carry one thing away with you: the assumption we brought up. You believe, somewhere beneath everything, that real knowledge is what you check for yourself, and that trust is the lesser thing you settle for. Do not let it drop out of sight again, because everything we do from here will bear on it.
Here is what you have, an hour in. You came in worried about being deceived. You leave knowing the trouble is deeper and stranger than that — that the real damage the fakes do is not the falsehoods you accept but the true things you begin to discard, and that the obvious cure, doubt harder, is the very thing the damage is made of. And you have the buried assumption out in the open, where it can no longer govern your response without your consent. That is what tonight was for. Not a solution. A clear view of what you are actually standing in.
Next time we do the thing I have been promising and declining all night: we put the strongest possible case behind verify-everything. There is a philosopher, a very great one, who holds that the careful, checking, prove-it-first instinct you have been feeling is not an anxiety at all. It is wisdom. And he holds that the deepfake era is not the end of anything — it is a vindication, the world finally compelling you to do the checking you always should have done. I am going to make his case as strong as I possibly can, holding nothing back, because if we are to find out whether you can live by it, we must first want to.
But that is next time. Tonight ends where it began. The phone will ring again. Perhaps not your sister, perhaps not at eleven at night, but something will come to you secondhand and ask you to believe it or not, with real money or real trust at stake, and no way to be certain. I told you at the start that I would not give you a way to detect fakes, and I have kept that promise; and this is not coyness, not saving the best part for later. There is no test I could give you that a good enough lie could not pass. You knew that on the phone. I knew it with my wallet in my hand. So you will have to answer it anyway, as I did — not certain, simply deciding.
The only thing I have given you tonight is a question to take to the phone. When it rings, and you decide whether to believe the voice or not, ask yourself how much of what you are doing is really checking, and how much of it is something else — something you do not yet have a name for. Do not settle it. Because whatever that something is, it is the thing we may have been misnaming all along, and starting next time we are going to find out what it actually is.
Thanks for listening. I'll see you next time.