‘It’s great … except when it’s not’: RishGPT reveals his insights on AI

We are now into uncharted ontological and epistemological territory. Deep into the metaphysical world. How would you know if Rishi Sunak was giving a speech on artificial intelligence? Could you ever really know that it was him? Or even if there really is a him? And could you differentiate the words from those generated by ChatGPT on an off-day?

Artificial intelligence has progressed exponentially in recent years. As has artificial stupidity. So much so that it’s often hard to differentiate between the two. Only seven years ago, the Conservative party gave us Theresa May. The original 1980s Amstrad version of a prime minister. The Maybot. Primitive malware that could only repeat a few phrases in a staccato bark. Even so, it took the Tories three years to realise she was completely ineffective and junk her. They wanted to believe and finally had to accept the evidence that had been staring them in the face.

Now we’ve got Sunak. Or RishGPT, as he likes to be known. More avatar than human. Albeit a rather more sophisticated shapeshifter. He looks and sounds far more plausible than the Maybot. His sentences even conform to basic grammar. Where things rather fall apart is at the level of meaning. All too often he merely talks in computer-generated soundbites. So much so that there is an almost total disconnect between what he thinks that he is saying and the words that come out of his mouth.

Nor is he helped by the complete lack of affect in his voice. Just a chilling, numbing monotone. AI powered by Valium. Or maybe he just doesn’t really care about anything all that much. Maybe he’s just not that curious about the world. His place in it is totally secure and that’s all that really matters to him. Everything else is merely incidental data. He’s just channelling his inner bored hedge fund manager as he casually informs an investor that he’s lost another five mill. But not to worry. These things all even out in the end. Sometimes.

But we are where we are. Stuck with RishGPT until the next election. So maybe it’s best not to get lost in the weeds of whether we have a prime minister who actually exists or not. Better to take his ChatGPT word for it. So it was off to the Royal Society to hear what he had to say about AI. The hottest ticket in town. Originally we were told the venue was too small for sketch writers. AI works best in a cupboard, apparently. Then late in the day we were told that two more seats had been found. We arrived to find ourselves in a vast, opulent room. With room for another 100 if necessary. No matter.

It wasn’t immediately clear why Sunak had chosen to devote a Thursday morning to a speech on AI. It’s not exactly a subject that is going to win him the next election. But I guess it’s one that interests him more than fixing the NHS, cutting inflation or dealing with the cost of living. And one of the bonuses of being a prime minister – even an interim prime minister – is that you get to sound off about whatever you like and journalists, friendly Tory MPs and a handful of bemused guests have to pay attention.

Call it the legacy issue. RishGPT is never going to be remembered as a great prime minister. More that bloke who was shuffled into No 10 after Liz Truss completely fucked things up. The man to manage the inevitable decline. But this way, there’s a chance that a few – we happy few – may later remember him as the leader who said a few rather obvious truths about AI.

Sunak began in a rather upbeat mood, though delivered in a downbeat way. There was no way anyone was going to get too excited. AI was great. It could do wonderful things, like help diagnose cancer. But then came the caveat. “I say this not because it is easy, but because it’s the right thing to do,” he murmured. He would repeat this line at least four times in 15 minutes. He clearly means it. It’s become his defence mechanism. He needs to believe the reason the country doesn’t like him is because he keeps telling it difficult things. The truth is prosaic. The country doesn’t like him because they don’t trust him and thinks he’s fairly useless.

skip past newsletter promotion

Our morning email breaks down the key stories of the day, telling you what’s happening and why it matters

Privacy Notice: Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Privacy Policy. We use Google reCaptcha to protect our website and the Google Privacy Policy and Terms of Service apply.

after newsletter promotion

On we went to the dangers of AI. Terrorism, cyber groups, paedophiles, pandemics, nuclear war. The apocalypse. And even though no one quite understood what they were doing, everyone was going to carry on anyway. Because the first one to stop was a sissy. And why let the Chinese work out how to unintentionally end the world when we could do it ourselves. It’s what made Britain great. We were all destined to play perpetual catchup. Only really understanding the capabilities of AI when the technologies had moved on to another level that we didn’t understand. The world’s dumbest arms race.

“But I don’t want to alarm you,” RishGPT said. Having in effect announced the end of the world. Because he was going to have a summit in the UK next week and establish an institute dedicated to failing to understand AI. And he would invite the Chinese to come, though he didn’t know whether they would show up or not. Or whether anyone outside the room really cared what he had to say about AI. The US and China are light years ahead of us anyway. So on the bright side, he would like to end by adding that AI was being used to clamp down on benefits fraud. Phew. Let’s bang up some claimants in prisons that are already too full while we wait for Armageddon.

There was then time for questions. Wouldn’t it be best to call a moratorium on AI? Definitely not. You couldn’t halt progress. Even if it killed you. Besides, it was never too late to hope for a new and improved RishGPT 2.0. One that was not totally ineffectual. And was he worried about the use of AI disinformation in the coming election? Sunak nodded earnestly. That would be terrible. The idea that democracy could be subverted was unthinkable. He would be devastated if the Tory party started running disinformation campaigns claiming that crime was down, waiting lists were falling, 40 hospitals were being built and that refugees were to blame for everything. That really would be the end of the world.

Leave a Comment