My AI prediction

Whenever people find out that I’m a speech writer, they ask me how AI is affecting my business. I always tell them that it doesn’t, and the reason for this is firstly because the people who are interested in hiring a speech writer are normally looking for that personal touch, the human element. They’re looking to hire an expert, and so they’re not really interested in AI. It doesn’t enter into their minds.

I’m not saying that there aren’t people out there who use AI for speeches. Of course there are. But the overlap between those people and people who hire speechwriters like me, it’s actually very, very small.

So what I’m about to say is not said with any prejudice. AI and I are not enemies.

Anyway, here’s my AI prediction: In the next 5 to 10 years, a lot of people in the mediocre level will be using AI to write their speeches. A lot of people already are. But what this will mean is an improvement in the profiles of those who don’t.

Those who use AI, the more that they use it, the less they’ll have a grasp on their subject, the less capable they will be speaking on the fly. And, in turn, the more capable those who don’t use AI will be, the more confident they’ll be, the more they’ll stand out.

The reason for this is because AI is an extreme form of the problem of delegation.

The Roman stoic philosopher Seneca talked about this 2000 years ago when he described a very wealthy man who had a large number of slaves, and he assigned each one a particular subject. It might be literature, mathematics, history, whatever. And anytime he wanted to know something about any of those subjects, or any time he needed to make a decision, he simply asked the relevant slave.

Seneca points out that the man has just simply hired other people to do his thinking for him. And that’s the problem of delegation at its core. Whenever we delegate anybody or anything to do a task for us, we shut off our own experience and our own opportunity for thinking within that task. Obviously, up to a point, that’s a good thing because the payment that we get back in terms of time or freedom, whatever it might be, is worth more to us than that opportunity. But only to a certain point.

But beyond that point, we lose our own thinking abilities, and our own grasp of what we’re saying. We become puppets.

Let me explain.

When I’m researching for a speech – and I spend about twice as long researching as I do writing – I never use any technology to summarise to get the key points. I don’t use AI. I go through and I read everything myself. For one single speech, for example, I had over 400 pages of printed material printed in addition to my electronic sources that I was reading. But why do I do all this? Not because I’m anti-technology. I do it because I’m giving myself an immersion in the subject matter, and this allows for the ‘magic of ideas’.

What do I mean by the magic of ideas? Well, it’s simple. It’s cross-pollination between a random word or phrase in one thing I’m reading, and an idea in some other bit of research. Often completely unrelated, no obvious connection. But then later on when I’m reading something else, I see a connection. Something else I read connects to that comment that I read in that first paper, and suddenly, as they say, it’s like a spark or a light bulb goes on. And in that moment, I have inspiration, I have idea. An idea that would not have come in any other form.

Now, if I had paid a person, a human, to summarise each of these documents, I would’ve got his or her interpretation, and so I wouldn’t have got those same connections. I had to be in that immersed experience. Without it, inspiration doesn’t come.

Now, I’m just giving an example of hiring a human writer. Obviously that summary that he or she gives me, I can use that to do the same thing, but every time we distil it further we reduce the chances, reduce the possibility for that to happen. That’s the problem.

Now, when you are going to use AI, you’re like that wealthy man who’s paying people to think for him. Because what’s happening here is essentially you’re never seeing the raw material. And that means that the AI is doing all the collating. It’s kind of like an executive summary –you’ve never read the actual report, only the summary.

Of course, you might as if that isn’t the same as hiring a human speech writer. And there are some similarities, but there’s a crucial difference here.

First, when you hire a good human speech writer, the first thing they’re going to do is ask you questions. They’re not going to go away and research. They’re not going to give you a draft or summary or anything like that. They’re going to ask you questions. In fact, I have a document that I give my new clients, which has 59 questions that I want them to answer before I start writing and research.

Why? Because these questions help me to understand the audience, understand the message, and understand the speaker. So right away we have two differences: a human writer (if they’re good) will ask questions. And, second, they know which questions to ask. I know this from experience, I’m going to ask different questions depending on who I’m talking to.

AI just doesn’t do that. It doesn’t come to you and say, ‘Hey, you want me to write a speech? That’s fine. Here’s a list of 50 questions that I need you to answer so that I can truly understand what you’re looking for, where you want to go, how you want this to turn out.’

AI won’t do that – but a human writer will.

And those questions are crucial.

Remember the criticism that Seneca made of that wealthy man was that he hired people to do his thinking for him.

But if a good speech writer comes back with questions, then you are still doing some of the thinking, which means more of your personality would be in there.

You might say, ‘But didn’t I hire a speechwriter so I don’t have to do all the work?’

Sure, and even when you have to answer these questions, it still takes less time and effort than having to do the research and the writing and the thinking all by yourself. So you still save time, but hiring a speechwriter is also about wanting the speech to be the best it can be.

Anyway, going back to the problem of delegation, when we hire a human speech writer, we shut off part of that immersion opportunity I spoke of earlier. But what we get in return is an expert who asks us questions that we wouldn’t even have thought of asking ourselves, somebody who will give back and forth with you on these subjects. And that back and forth is a crucial element.

Let’s say you work with me. I start off by asking you questions, and those questions already make a difference between a human writer and AI. So I ask you those questions, you have to do a bit of thinking, and then you answer those questions.

Once you answer them, I go away and do the research with a clear picture of what you’re thinking. At the same time you have a clearer picture of what you’re thinking, where you want to go, and so on.

Then when it comes to the first draft, we see another difference between AI and humans. If  you got that draft from AI, you have two options. You can either take the draft as it isand work with it, edit it yourself, or you can come back and try and get AI to rework it.

That first option is the same for a human or AI writer. You can say “Thank you. I’m going to take it from here” if you want.

But the second option is very different. If you come back to me, we can get on the phone and talk through what you’re thinking here and there, and I can make those changes. Sometimes they’re tiny changes, sometimes they’re big changes, but I’m interpreting them and discussing them with you. But AI can’t make those judgement calls.

So if you go back to AI to do the rewriting, you have to go through and give it specific instructions. But if you’re doing that, then you’ve actually chosen option one, which is to do the editing yourself.

And this is the point: if you trust it too much, you become like the wealthy man who had his slaves do the thinking for him. But since you’re not involved in the thinking, you become too dependent on it. And that’s how you end up with situations like the lawyer who used AI for legal research and had phantom cases because he just trusted it to be right.

But if you don’t trust it, then you have to do a lot of the work yourself – which means you’d have been better off hiring a human writer you do trust. If you’re that kind of person, AI is not offering a solution for you, because you need somebody who can think, and help you think. Someone to help you put your best foot forward every time you speak. And – crucially – you want to be confident that you are across the subject yourself. Every good speaker should be. AI just can’t help you do that.

So this is my prediction. In the next 5-10 years, we’ll see two groups of speakers, those who rely on AI and those who don’t. the first group will be increasingly incapable of speaking on the spot, responding on the spot, making good cases for any point that they wish to make unless it’s already been prepared, and in general won’t be confidently across their subjects.

The second group will be those who are confident, able to speak across the subject, able to answer the good, hard questions. And the more that this divide grows, the more that confident second group will become. The more popular they’ll be. The more they’ll stand out.

That’s all, folks.

Alexander

P.S. If you want complete confidence that your next talk will be the best version it can be – and one that helps you clarify your own position – then sent me an email at [email protected] and let’s talk!

P.P.S. The other option is I could be completely wrong in my prediction – but hey, it was fun to think about!