Strong AI
Strong AI is the thesis that an appropriately programmed computer would not merely simulate a mind but literally have one, with real understanding and mental states, in the same sense that people do. The contrasting position, weak AI, holds that computers are powerful tools for studying and modelling the mind, and that a program simulating thought is no more thinking than a simulated storm is wet. The distinction concerns what a system is, not how well it performs, which is why it cuts across questions of capability and is so often misapplied in popular discussion.
The terms come from the philosopher John Searle, who introduced them in his 1980 article "Minds, Brains, and Programs" in the journal Behavioral and Brain Sciences. Searle used strong AI to name the claim that the right program, running on any suitable hardware, constitutes a mind, so that the study of cognition reduces to the study of programs, while weak AI is the claim that computers merely help test psychological theories. He directed his argument at the view, held in various forms by researchers in the symbolic tradition and by functionalist philosophers, that mind is to brain as program is to hardware. The article appeared together with dozens of commentaries from cognitive scientists and philosophers, and the debate it set off is still active.
His central argument is the Chinese room: a person who understands no Chinese sits in a room with a rulebook for manipulating Chinese symbols, and by following the rules produces answers that convince outside observers that the room understands Chinese. Searle's point is that the person, running the program by hand, still does not understand a word, and that the symbol manipulation alone, which is all a program does, cannot suffice for understanding. He distinguishes syntax (formal symbol shuffling) from semantics (meaning) and argues that programs are purely syntactic. Passing a behavioural test such as the Turing test would, on this view, show nothing about whether understanding is present, so the argument is a direct challenge to behaviourist readings of that test.
The replies are as instructive as the argument. The systems reply says that understanding belongs to the whole system of person, rulebook and workspace, not to the person; Searle answers that the person could memorize the rules and still understand nothing. The robot reply says that grounding symbols in perception and action supplies meaning; Searle answers that adding more symbolic inputs changes nothing. The brain simulator reply asks what happens if the program mimics a brain neuron by neuron. Searle's own positive view, sometimes called biological naturalism, holds that minds are caused by specific biological processes in brains and that a computer could think only if it reproduced those causal powers, not by running the right program. Others, including Roger Penrose from a Gödelian direction, and Hubert Dreyfus from a phenomenological one in his 1972 book What Computers Can't Do, criticized strong AI on different grounds.
Strong AI is related to artificial general intelligence, but the two are not the same thing. AGI, a term popularized in the early 2000s, refers to a system able to perform a wide range of intellectual tasks at roughly human level, a claim about capability that could in principle be met without any inner experience. Strong AI is a claim about mind. A system could in principle be generally capable yet, on Searle's view, understand nothing, while a hypothetical conscious machine might be narrow. In everyday usage, "strong AI" is often wrongly taken to mean powerful or superhuman AI, and Searle's original meaning is lost.
The debate returned with large language models, whose fluent conversation revives the Chinese room question. Their defenders point to emergent competence and the systems reply; critics call them statistical text predictors that manipulate form without access to meaning. A common misconception is that Searle argued machines can never think; he allowed that a machine, the brain being one, can think, and denied only that running a program is sufficient. A second is that the thought experiment settled the matter, when philosophers remain divided over it.
Read strong AI as a thesis about what a machine is rather than what it can do: the claim that the right program is itself a mind, which Searle's Chinese room attacked by separating symbol manipulation from understanding, and which the capabilities of today's models make both more tempting and no easier to settle.