AGI, or artificial general intelligence, refers to an AI able to match humans on most intellectual tasks, not just on one specific task. There is no shared definition: OpenAI's charter speaks of highly autonomous systems that outperform humans at most economically valuable work. Superintelligence goes further: an AI that would clearly surpass humans in almost every field. In October 2025, the Future of Life Institute published a statement, signed among others by Geoffrey Hinton, Yoshua Bengio and Steve Wozniak, calling for a ban on its development until there is both a scientific consensus on its safety and public buy-in.
In practice at Gensai
Gensai starts from what models can do today: every project answers a precise, measurable need, without waiting for general intelligence.
Should you wait for AGI before starting a project?
No. Today's models already handle targeted tasks very well: answering from documents, sorting requests, drafting a first version, holding a simple phone conversation. A well-built project relies on these verifiable capabilities and benefits from model progress version after version.
Common mistake
Treating AGI announcements as a roadmap. There is neither a common definition nor a recognised test to say it has been reached, and the dates put forward vary widely from one player to another.
Same theme
Definition written with AI assistance, checked by Sébastien Lecocq, co-founder of Gensai.
Going beyond the definition?
From concept to project: Gensai builds custom, sovereign, GDPR-compliant AI solutions.
Related services: Custom AI solutions, AI agents