Should governments really be using AI to remake the state?

What is Artificial intelligence? Scientists are a question that fought on the dawn of computer scientists in the 1950s Alan Turing Ask: “Can machines think?” Now the great language models like Chatgpt (LLMS) are released in the world, a response is never more pressured.

There has been extensive use, social norms on these new tools are evolving rapidly. They should use students Type attempts? They will do Replace your therapist? And the Government of Turbozigraphy?

This last question is made in the US and the United Kingdom. According to the new Trump Administration, Elon Musk’s government efficiency department (DOGE) deletes the federal staff and directs GSAI, GSAI, GSAI, who remain. Meanwhile, Keir Starmer calls the UK Prime Minister, the “Gold Opportunity” that can help redap the situation.

Definitely there is a government work that can benefit from automation, but is the appropriate tool for LLMS job? Part of the problem is still agreeing to us. This week was proved appropriately, when New scientist Freedom of information (Foi) used laws Get Peter Kyle’s reunionSecretary of State for Science, Innovation and Technology in the UK. Politicians, data privacy experts and journalists – do not surprise us that request was given that request to Google Search history, say, would generally be rejected.

The UK government released the UK says Chatgpt requires the use of civil functions in ministry interviews, via email or whatsapp, which are both subject to the laws of Foi. Keyen’s interaction has no reliability with Chatgt to form a serious policy in AI – one of which the podcast had to appear. However, having been given to Foi requests, some of the government seems to believe that Alas can speak like a man.

So New scientist It is very widespread, current LLMs are not meaningful and responsible for any sense Unspecified mistakes As they offer useful advice. Even their answers will also be reflect the innate naughty information that swallow.

In fact, many AIKA scientists are increasingly llms that are not the goal of the general purpose (Agi) in general, which man can do or overcome any man can do. For example, in the end AI Researcher SurveyAbout 76 percent of respondents said it was “unlikely” or “very difficult” that today’s approaches will be successful in achieving agi.

Instead, we may think about these Ais in a new way. Writing in the magazine Science This weekA team of researchers said that “they are not primarily to see smart agents, but as a new type of cultural and social technology, other humans have accumulated to human information.” Researchers LLMs are “past technologies” for writing, printing, markets and representative representatives due to democracy and changed the way of processing and processing.

In this way the frame, the answers to many questions become clearer. Can governments use LLMs to increase efficiency? Almost definitely, but when people understand only people and limitations. Interaction Interaction should be subject to freedom of information laws? Valve designed to provide “safe space” for internal deliberation. And, Turing asked, do machines believe? No. Not yet.