In today’s rapidly evolving world, artificial intelligence is becoming a hot topic, especially when it comes to its role in government operations. This conversation really took off after a UK minister was found using ChatGPT for advice, sparking curiosity about AI’s potential in government and how smart these systems really are.
The Trump administration is looking into AI to make government functions more efficient. Leading the charge is Elon Musk’s Department of Government Efficiency (DOGE), which is exploring the idea of replacing federal workers with AI chatbots like GSAi. Across the pond, UK Prime Minister Keir Starmer sees AI as a “golden opportunity” to revolutionize state functions.
The debate around artificial intelligence isn’t new. It goes back to the 1950s when Alan Turing asked, “Can machines think?” Today, with advanced language models like ChatGPT, understanding their role is more crucial than ever. These AI tools are widely used, yet we’re still figuring out their social implications. Should students use them for essays? Can they replace therapists? And importantly, can they make government operations more efficient?
Automation could certainly improve some government processes, but whether language models are the right fit is still up in the air. A recent freedom of information (FOI) request showed UK Secretary of State Peter Kyle’s interactions with ChatGPT, which were treated like ministerial emails or WhatsApp chats. Although these interactions weren’t policy-focused, their disclosure under FOI laws suggests a belief in AI’s conversational utility, raising some eyebrows.
It’s important to remember that current language models aren’t truly intelligent. They can generate both useful suggestions and convincing inaccuracies, and they might carry biases from their training data. Many AI researchers are skeptical that these models will achieve artificial general intelligence (AGI), where machines would match or exceed human capabilities. A survey found that 76% of AI experts think current AI methods are unlikely to reach AGI.
Recent research in the Science journal describes AI as a “cultural and social technology,” similar to transformative inventions like writing and printing. This perspective helps clarify how language models might be used in government: they can boost efficiency if used by knowledgeable people. Plus, interactions with chatbots might fall under FOI laws, but the spaces where ministers deliberate should remain confidential.
So, while AI can certainly assist in government processes, it’s not yet capable of autonomous thinking like Turing envisioned.
