What is AI, and what are its main limitations in a government context?
Answered
Learn what AI can and cannot do before using it. Systems currently lack reasoning and contextual awareness, are not guaranteed to be accurate, and — for generative models — produce probabilistic, human-like output without understanding meaning.
From the guidance
Primary (how) AI Playbook for the UK Government
AI systems currently lack reasoning and contextual awareness and their limitations vary depending on the tools you use and the context in which they operate. AI systems are also not guaranteed to be accurate.
Section: Principle 1: You know what AI is and what its limitations are
Read this in AI Playbook for the UK Government (opens in new tab)
Secondary (normative) AI Insights: Generative AI
these models do not understand the content or meaning of the words beyond how an input set of numbers may most likely translate into an output set of numbers, and subsequently into streams of words. This means that there are limitations on what these models can reasonably do. You must consider that these are probabilistic models, and, despite producing human-like output, they are not at all sentient.
Section: How generative AI works