Glossary
What is strong AI (AGI)?
In short
Strong AI, also called artificial general intelligence (AGI), is a hypothetical system that can do any intellectual task at the level of a human or higher.
AI basics · AGI
MMOOZZGG.com editorsChecked: 1 min read
For a child
Your calculator counts faster than anyone in your class, but it cannot draw, play ball or comfort a friend. It is good at only one thing. Scientists dream of a machine that could do almost everything a person can: learn new things, invent, understand jokes. There is no such machine yet. People argue about when it will appear, and whether it will appear at all.
In plain words
Today’s AI systems are strong at separate jobs: one writes texts, another recognizes faces, a third plays chess. They are called weak or narrow AI. Strong AI is AI that carries knowledge from one area to another by itself, learns unfamiliar work without retraining and reasons as well as a person. There is no single definition: some see breadth of skills as the main sign, others independence, others consciousness. So the claims “AGI is already here” and “AGI is decades away” often differ only in words.
For business
For a company, strong AI is not a product yet but a topic of forecasts and investment. You cannot buy it, and you have to make decisions now, on the models that exist. A practical approach: build processes so that you can swap the model for a new one without rebuilding the whole system, and do not pay for a vendor’s promises. If a contractor sells “almost AGI”, ask them to show results on your tasks and in your numbers. The risks come from inflated expectations: projects planned around a future intelligence miss deadlines and budgets. A review of the market of solutions and rollouts shows what works today.
For a specialist
The philosopher John Searle described “strong AI” in his 1980 paper “Minds, Brains, and Programs”: in this view, a properly programmed computer does not simulate a mind but has one. Later the meaning moved toward artificial general intelligence, that is, the breadth and level of abilities, not consciousness. The 2018 OpenAI charter describes AGI as highly autonomous systems that outperform humans at most economically valuable work. Morris and co-authors (2023, ICML 2024) proposed a scale with two axes: the performance level (Emerging, Competent, Expert, Exceptional, Superhuman) and breadth (narrow or general AI), with autonomy treated separately. The authors placed chat models such as ChatGPT at the first stage of general AI; higher stages were considered not reached at the time of publication. Measurement relies on test sets (benchmarks), but there is no single criterion by which AGI is declared achieved.
Example
A vendor promises a director “an employee with general intelligence” who will replace the analytics department. The director asks for a pilot on one task, the quarterly sales summary, and sees that the system copes, but without a human check it makes mistakes in data exported from 1C, the accounting software.
Related terms
Questions and answers
Sources
Not sure whether your company needs this?
AI solutions map