AI is the broad field of systems that perform intelligent tasks. AGI and ASI describe proposed broader capabilities; they are not synonyms for a newly released language model.
What do the terms mean?
An AI system can be excellent at a specific activity such as translation without being reliable at everything. Artificial general intelligence (AGI) refers to broadly capable intelligence; definitions and ways to measure it differ. Artificial superintelligence (ASI), sometimes shortened to SI in this context, refers to a proposed system surpassing human capability across a broad range of intellectual work.
How should you assess a claim?
Ask which tasks were measured, what information the system received, whether tools or people helped, and how often it failed. Success on a selected exam does not demonstrate competence on every unfamiliar task. A company's label is not an independent evaluation.
For a learner, the practical priority is checking what an available system can do reliably. You do not need a forecast for AGI or ASI to learn prompting, retrieval or evaluation.
Try it yourself
Write three columns: demonstrated result, proposed capability, and unknown. Classify a claim you have seen using its actual evidence. If evidence is missing, record that instead of treating a prediction as a fact.
Source and next step
Read IBM's explanation of ASI for terminology. Continue with model selection.