Revealed: How the UK tech secretary uses ChatGPT for policy advice
newscientist.com
newscientist.com
I'm very surprised this is subject to FOIA. I'll bet he learns to disable memory and uses incognito chats now though.
The process of doing that clarified my thoughts and arguments so much that I never wound up having to send them -- I'd already made the right points on Slack and in meetings, and a compromise was achieved.
As you dig down it's often so hilariously useless it sparks my decaying intelligence to offer up a gem. Less fruitful than talking to a Junior Engineer but more useful than a physical rubber duck.
Like how engineers claim they are not losing skills because they are using programming tool assistants, I think there is an over-confidence thing at play here, where the user of an AI assistant doesn’t want to admit to themselves that they are also being shaped by it.
You are still being subtly influenced by something, that’s the whole issue at play here. OpenAI has incredible power in influencing the subconscious of so many people, and I think thats the thing to question here. I am not entirely against it, but we need to be honest with ourselves with how malleable our minds are and influenced we are by what we consume.
Concerns about bias are bikeshedding. I am much more concerned about the systemic bias of race than I am about OpenAI trying to psyop people(I'm not concerned about race).
I'm sure if he wasn't already aware, he's been briefed by his SpAd or civil servant (possibly via the security services) not to disclose any sensitive information.
It's one thing when a developer asks an LLM for programming advice, and the LLM has been instructed to prioritize answers suggesting React. In my opinion, it's quite another when a politician asks an LLM for policy advice, and the provider has the ability to make similar prioritizations.
All I see him asking for action on is which podcast to appear on. The rest is just run-of-the-mill research.
Isn't this true of every information source?
Problem is that the owner of the AI service has power over the model and thus can influence government decisions.
Maybe we should require that government use of AI should be required to use open models only.
Or, you know, consulting experts.
Sometimes they go to the other extreme and think experts are gods, and do not question the advice, check whether other experts would disagree with it etc.
Ask it about something you don't know: you'll learn something new.
Ask it about something you do know: you'll raise your eyebrow, "huh that's not quite right..."
The main problem is that it is easier to sound competent and less effort to spam. Like, the competence and effort filter is gone.
>The Globalist Wall Street Journal has no idea what they are doing or saying. They are owned by the polluted thinking of the European Union, which was formed for the primary purpose of “screwing” the United States of America... (Truth Social today)
It's completely bonkers that people keep saying stuff like this, virtually implying that we didn't have a functional society before LLMs.
LLMs are tools that generate word salads that sound compelling. They are not research aids and they can't help you understand things better than the plethora of tools we already developed over the centuries can.
So no, the alternative is policy makers doing what policy makers have always done: research, polling, reading, talking, and more research.
Now you'll likely come back and say "but X policy was bad and it was because it was poorly researched". Absolutely! Yes, bad policies exist. Poorly researched policies exist. Poorly implemented policies exist.
Good policies researched well also exist. And the bad ones aren't going away because of the magical word generator.
Seriously. Stop using LLMs to "help" you do stuff you already know how to do.
In the United States, lobbyists have approached lawmakers with pre-written legislation and pitched it to them.
I would rather have lawmakers asking LLMs to explain concepts to them than the absolute mess we had back when a House committee was trying to evaluate SOPA with absolutely zero domain knowledge:
I addressed that: bad policies and decisions don't go away just because of LLMs. They will always exist, it's just that now they'll exist at the same time as people gaslighting themselves into believing that LLMs are helping to eliminate them.
> Now you'll likely come back and say "but X policy was bad and it was because it was poorly researched". Absolutely! Yes, bad policies exist. Poorly researched policies exist. Poorly implemented policies exist.
> Good policies researched well also exist. And the bad ones aren't going away because of the magical word generator.
You will notice that the US is gearing up for additional warfare with the help of AI.
Why?
Given the current state of LLMs it would only be giving advice but even so would probably give better advice than some of the idiots that get elected. Also maybe more importantly it would be a known quantity whereas humans can be deceptive.
https://en.wikipedia.org/wiki/List_of_assigned_/8_IPv4_addre...
Here is the nginx proxy rule so that requests from the UK government go to the right place:
if ($remote_addr = 25.0.0.0/8) {
set $backend_pool grok_with_special_information_for_really_smart_uk_tech_ministers;
}
else {
set $backend_pool grok_for_poor_and_stupid_uk_people;
}
This was generated with help from Claude, so please verify your own sources.
The most insidious part of this, is it gives us the feeling that we are still in control.
When in reality, a few extremely powerful and extremely wealthy individuals are in a position to dramatically shape and shift policy.
15 years ago Google was still pretty damn good. Today most people agree that it’s departed from its original goals and in its efforts to make money actively censors and shapes the results it serves.
The AI systems people use to make decisions will make this shift once they have a lock on the market.
It’s wild to see the creation of a technological equivalent to the medieval Catholic Church. an entity that represents itself as benevolent to the masses, but in reality, exploits them and maintains a stranglehold over the political ruling class.
But I don't recall last time I had to click page 2 of their search.
It would be worrying if the report was that Kyle based some significant part of his decision-making on AI chatbots, but no-one here seems to be claiming that's the case. As it is, given the article only writes about two significant examples and a bunch of definitions, it sounds like he's not been using it much at all.
Certainly, information leakage from search activities of any kind by high government officials is a cause for concern (as are travel patterns etc.).
You could identify which world leaders are using ChatGPT pretty easily.
The courts use MS Teams for remote hearings - not as bad, but I do not like it either.
The FOI request would have been sent to the government not Open AI, so I do not think Open AI would have needed to identify the individual.
Sure I already suspected such a thing given what has been happening for the past decades but they still manage to surprise me.