Ulf Kristersson, Prime Minister of Sweden and leader of the country’s conservative coalition, He has publicly recognized that consults artificial intelligence tools such as Chatgpt and Mishwner (Mistral European chatbot) to obtain “second opinions” in its government decisions. A statement that has unleashed a political and academic storm about the risks of using AI in such critical sectors as public and citizen management.
In the focus. In An interview With the Dagens Industri Economic newspaper, Kristersson not only admitted their personal use of these tools, but also revealed that their collaborators also employ them in their daily work. “I use it quite often, even if it is to obtain a second opinion” said the prime minister. Apparently, Kristersson claims to use the tools to ask things like “What have others done? Or should we think just the opposite?”, He explained about the questions he asked.
A devastating reaction. The Actonbladet newspaper accused directly to Kristersson having “fallen into the psychosis of the AI of the oligarchs.” Meanwhile, some experts have shown concern about security and democratic implications. Simone Fischer-Hübner, researcher and computer expert at the University of Karlstad, warned about the dangers of introducing sensitive information in Chatgpt: “You have to be very careful.”
The warning makes sense, since although the company has security and privacy measures, any conversation we have with the chatbot ends on OpenAi servers. Therefore, it is not a very safe approach, especially if we talk about such a critical use as in the political dome.
Beyond security. Virginia Dignum, professor of artificial intelligence responsible at the University of Umeå, assures that AI cannot offer significant opinions about political ideas, since it simply reflects the biases of those who developed it. “The more it depends on AI for simple things, the greater the risk of excess confidence in the system. It is a slippery slope,” he warned. And he finished with a phrase that has turned the world around in networks and media: “We do not vote for Chatgpt.”
A defense that does not convince. Kristerson spokesman Tom Samuelsson tried to minimize controversy ensuring that the prime minister does not handle sensitive information through these platforms and that he uses them “more as a general reference.” However, Jakob Ohlsson, an expert in AI, Point out that even seemingly harmless information can reveal patterns of government strategic thinking to technology companies in charge of this type of systems.
It is these types of examples that show us The great adoption that have had these types of systems, regardless of the sector in which they work.
Cover image | Solen Feyissa and Anders Wiklund/AP
In Xataka | OpenAi has just kick the AI board. Your new model is free and can be downloaded to use from the laptop
GIPHY App Key not set. Please check settings