ChatGPT, political advisor to the British government

Transparency laws have achieved an unusual milestone: the UK's Technology Minister, Peter Kyle, has had to reveal his ChatGPT usage history. Among the searches that have emerged, one stands out in which he asked Kyle for advice on which podcasts he should appear on and also why the adoption of artificial intelligence in the country was progressing so slowly. Coincidentally, or not, this week British Prime Minister Keir Starmer asserted that his government should make more intensive use of AI to increase its efficiency and productivity. As I am more integrated than apocalyptic, according to Umberto Eco's old division, I am not against the adoption of these tools as a support for designing policies. However, it is essential to be aware that they have their biases and that they operate with opaque algorithms (even open-source ones offer answers that are difficult to trace), so we will have to be very careful to ensure that these tools are used effectively. spin doctors pocketbooks do not end up governing a territory with ideas inspired by their powerful owners, the neo-feudal lords of thebig tech.

Asking for brainstorming sessions is one of the basic applications that AI can adequately fulfill. But we must be aware that these will always be secondhand ideas, extracted from some content from who knows where on the farthest reaches of the internet. And they will hardly be neutral. If the minister in question is a truly expert and rigorous person, they should know how to separate the wheat from the chaff and use the response obtained as leverage to articulate their own policy. If they are lazy or one of those multi-portfolio professionals who jump from one ministerial flower to another, there is a greater danger that they will end up uncritically buying into a policy designed... by whom? And, above all, for whom?