What is AGI, and should we worry about AI risks?

Artificial general intelligence (AGI) means an AI system that can match or exceed humans at almost any intellectual task, not just one narrow job. Today's AI, however impressive, is not generally accepted as AGI, although leading systems are becoming capable across a widening range of tasks.

When might it come? Experts genuinely disagree. Some leaders of AI companies say it could arrive within years. Many researchers think it will take much longer or question whether current methods will get there.

What risks are discussed?

• Near-term: misinformation, deepfakes, scams, bias, job disruption, privacy loss and cyberattacks.
• Misuse: AI helping bad actors with dangerous weapons or large-scale hacking.
• Concentration of power in a few companies or governments.
• Loss of control: a longer-term concern that very capable AI systems might pursue goals that conflict with human interests.

What is being done: AI safety research, testing of models before release, government AI safety institutes in several countries (India has proposed one), and international summits. These began at Bletchley Park (UK) in 2023, continued in Seoul and Paris, and India hosted the AI Impact Summit in 2026.

A balanced view: take the risks seriously without panic, and support sensible rules alongside innovation.