Prime Ministerial Insights - CONFIDENTIAL
It’s going to be a good week!
My wife, Akshata, reminded me we are hosting the big AI Safety Summit next week. I had almost forgotten about it to be honest. (NOT!!! I’ve been giddy as a kipper with excitement for weeks - finally something that can’t crash around my ears and will see me take my place on the global leadership stage!)
Akshata reminded me that this summit really is a big deal. What with the UK being the birthplace of AI thanks to Alan Turing’s test thing. And everyone, but everyone is going to come. We have the Americans, Europeans, big tech companies, even China is coming. This is going to be big for business, big for Britain and big for the Conservative Party. And big for votes, lots of votes.
We need AI right now. It will sort out the asylum backlog and the NHS queues and reduce taxes for everyone. Except maybe for most of the voters, but we don’t have to tell them about that.
All hell breaks loose!
Damn it! The Canadians just published their new list of AI threats. So much for the Commonwealth of nations! And Justin’s got a diary clash, so he doesn’t get to take part in the Bletchley spotlight! His loss!
The Canadians seem to be blathering on about “Autonomous AI”. That’s far too many syllables. I could never put it in a speech. I would definitely lose votes.
“What about this autonomous thing,” I ask Akshata. “How the heck am I going to explain it? It’s too many syllables.”
“You’re right,” she says. “Keep it simple. It's like the Tory rank and file. Just when you get everything under control and are doing all the right things, they throw a curve ball. Those researchers are talking about AI agents, like Siri and Alexa, only they collaborate and conspire against you. Better to call it Agentic AI.”
“You mean like Alexa and Siri gang up on us,” I ask. “But they are our friends, and they are good for business. Apple is worth more than the whole FTSE. We want them in London for the AI Safety Summit.”
So, I realise we must include this new thing at the Summit. It's going to protect jobs and make the world a better place. And here in the UK, we are doing a much better job at not regulating AI, so everyone is going to come here.
More taxes, more jobs, and better opportunities for everyone. Maybe Elon Musk will come? He could take over the abandoned EV battery factory and maybe make a right-driving version of his car again. We got a whole battery factory ready for him. Is he coming?
I better check in on him and make sure he knows he is invited. Isn’t he building an AI supercomputer?
Doesn't really matter if he doesn't come - everyone else is! His loss!
All hell breaks loose!!
Some of my advisors tell me some of the AI Safety Summit attendees might travel to Bletchley Park by rail.
“Sounds good to me,” I say. “We are a green country with lots of trains that definitely run on time and aren’t completely unreliable, when they’re not on strike. Sustainable rail, not like the US.”
“But it goes from Paddington Station!” they say.
“What’s wrong with that?” I ask. “We have a cute bear and a lot of Harry Potter stuff. Won’t those techies appreciate all those great things about the UK?”
“Yes, sir, but there’s that Alan Turing exhibit right outside the station.”
“Sounds good to me, isn’t he the father of AI? What could be wrong with that?” I fire back.
“It’s the other bit, sir, the apple and the gay thing.”
“We like gays,” I remind them. “We have Pride and all that diversity and inclusion stuff.”
“But sir, the cyanide apple thing, it’s bad optics.”
“Oh goodness,” I fire back. "Can’t we just cover it up for maintenance or cleaning?”
“Yes, sir, we will see what we can do.”
Problem solved. Besides, didn’t we pardon him or something?
We are all in on AI, including the LGBTQ alphabet people, so long as they don’t try to use the wrong toilets or change sports teams.
Anyway, why is anyone taking trains in the first place? What’s wrong with public transport, like a helicopter?
All hell breaks loose!!!
The Summit is still happening. We have the EU, the Chinese, the Americans and all the big tech companies - be lovely to catch up with Nick Clegg and reminisce about those good old Coalition days in government! The UK is going to be the leader in AI safety, and it’s going to be good for citizens, businesses and bankers, and if some Indian services companies do well out of it as well, that's all to the good. I imagine.
Then Deputy Prime Minister Oliver Dowden goes on the BBC - always a mistake - and tells them we are going to save jobs, improve efficiency and cut all the queues. All good stuff, long term decisions for a brighter future etc etc etc.
But now he has to get specific about the numbers. And they were bad numbers. He tells them benefits officers currently process about 11 applications per week, but AI is reducing this down to 200 minutes per application.
I did the math. Ollie was never good at math. That’s still 37 hours per week for 11 applications. And then how much is that AI going to cost us?
I call Ollie and tell him, “Don’t put the numbers in there. You will jinx everything. You were never good at numbers. Just keep it vague. That always worked at the Treasury and that never did us any harm!"
Then Akshata’s dad goes on Indian TV to tell Indians they all need to work 70 hours a week to catch up with China. Sounds good to me. I work at least that. Maybe we can bring that here?
Ollie does much better at dodging the Chinese questions. The BBC fellow kept asking him to get specific. But you never know with those Chinese. “We invited them, they might come,” which was honest.
We need to include the Chinese. They are AI leaders - and they're loaded. Besides, they have a lot of great ideas about how to frame all this AI policing - and they're loaded.
Then, of course, Madame has to throw in her tuppence worth! Liz Truss, my 49 day predecessor, starts banging on about the Chinese threat. Goodness gracious. All these fighting words like ‘autocratic,’ ‘state control,’ and ‘national security.’
Doesn’t she realize they make all the chips or at least package them up cheaply? Oh, that’s right, they don’t make the chips anymore since we cut them off, but somehow, they make all the gear using them cheaply. I mean. they could be dangerous, but they have really good ideas about how to frame all this surveillance stuff. We should definitely invite them.
I hope they come. We will see—still a few days to go. We won't mention the Uighurs. That’s a sore spot for them, and they are AI leaders. We need them at the Summit.
As for Liz, we shall just ignore her. She’s happy enough to deal with China when it comes to pork markets I seem to recall, lots of smug smirking on that front! I can’t waste any more time on her - I have to give a big and very important speech!
We are going to study AI regulation and spend £100 million on it. That’s a nice big round number. It shows we are not going to trust those big US companies to grade their own AI homework. And we are building AI supercomputers and quantum. Big data centers - bigger than a human brain. But we are not going to regulate it yet because we want them all to know that the UK is the place to invest.
AI is going to solve all our problems with the NHS, government services, reduce the number of boat people, improve the police, and save our eyes with AI-powered eye tests. It’s the place to invest. Come to Bletchley Park, we know how to not regulate stuff so you can turn a profit.
All hell breaks loose!!!!
Ursula von der Leyden says she may or may not come. Macron isn’t coming, which is really annoying as he’s the only one shorter than me in the photo opps! Word gets out that our so-called allies are now starting a new AI Safety Committee meeting as part of the United Nations. And they did not even invite us – traitors!
Oh yeah, they also promised not to do all of their AI regulation stuff just yet so they can catch up with the US and China by putting off the new EU Act for now. We might as well just join back up again. At least we still have US VP Kamala Harris still coming. (She is still coming, right?)
All hell breaks loose!!!!!
Oh why do I bother!?! Now my aides are telling me the US is coming out with a hundred-page report on AI safety today. And Old Joe, who’s too busy to come to my Summit, is going to be signing off on a really tough and enforceable AI Executive Order just to steal attention! Well, let him try. We have more smart people, and we have Cambridge and Oxford, and the Royal Society and Alan Turing. And we are giving away £100 million. Maybe we should write a 120-page report so we can show we have more words due to our world leadership? (Can we get ChatGPT to pull something together?) Hell, we can break it into three parts and make it a thousand pages. It’ll be something to read in the chopper on the way to Bletchley.
And now Ursula’s put out a press release deliberately talking up the EU’s role in the G7 international Guiding Principles & the voluntary Code of Conduct for Trustworthy AI, which is some kind of global thing, which is fine, but it’s not taking place at Bletchley Park, is it? So…well, let’s just see where her seat ends up being if she shows her face on Wednesday! if she doesn't turn up, her loss!
It’s going to be a long week.