Illustrating another weakness of AI…

https://techcrunch.com/2025/01/19/ai-isnt-very-good-at-history-new-paper-finds/

AI isn’t very good at history, new paper finds

… The results, which were presented last month at the high-profile AI conference NeurIPS, were disappointing, according to researchers affiliated with the Complexity Science Hub (CSH), a research institute based in Austria. The best-performing LLM was GPT-4 Turbo, but it only achieved about 46% accuracy — not much higher than random guessing. 

… “If you get told A and B 100 times, and C 1 time, and then get asked a question about C, you might just remember A and B and try to extrapolate from that,” del Rio-Chanona said.

The researchers also identified other trends, including that OpenAI and Llama models performed worse for certain regions like sub-Saharan Africa, suggesting potential biases in their training data.

(Related) I ask if this would work with an AI President Trump?

https://www.nytimes.com/2025/01/18/us/politics/cia-chatbot-technology.html?unlocked_article_code=1.qU4.ewJH.Y67Tn8psaZWH&smid=url-share

C.I.A.’s Chatbot Stands In for World Leaders

Understanding leaders around the world is one of the C.I.A.’s most important jobs. Teams of analysts comb through intelligence collected by spies and publicly available information to create profiles of leaders that can predict behaviors.

A chatbot powered by artificial intelligence now helps do that work.

Over the last two years, the Central Intelligence Agency has developed a tool that allows analysts to talk to virtual versions of foreign presidents and prime ministers, who answer back.

So where would the ‘power’ go? (Remember, AI can’t vote.)

https://www.theguardian.com/technology/2025/jan/19/ai-could-destroy-democracy-as-we-know-it

AI could destroy democracy as we know it

… Since the second Industrial Revolution in the late 19th century, the prevailing national political superstructure of industrial capitalism in the global north, apart from the interlude of European fascism, has been various forms of parliamentary democracy. Those structures developed, in large part, because organised labour could bargain with capital for a share of the wealth that human labour creates, and built political parties to represent working people’s interests. Indeed, labour-relations systems, based on freedom of association and collective bargaining, have been pillars of functioning democracies.

However, rather than creating more productive jobs, as some envisage, the AI revolution could entail a transformative reduction in work and employment that would remove capital’s reliance on human labour to produce surplus value and profit. If that leads to the demise of workers’ organisations and to further hollowing-out of the economic base of social democratic political parties, crucial questions arise. How will capital be kept under control and held to account? What will prevent the Musks of the world from achieving complete state capture? What mechanisms will be left to ensure some semblance of redistribution of the wealth created in AI value chains? And how far can incomes fall before levels of demand become unsustainably low?