AGI means Artificial General Intelligence, a type of AI capable of thought, learn, and practising knowledge in broad areas of activity at the level of a human being (or even higher).
In comparison to AI systems used nowadays, which are designed to accomplish a particular task, AGI would have the capacity to think, reason, and evolve in an expansive manner resembling a human mind.
The Comparison between AI and AGI
The previous AI systems are referred to as narrow AI today.
For example:
- ChatGPT is capable of text generation and responding to questions.
- Google Translate translates languages.
- AlphaGo was a master of the Go board game.
Such systems are high-powered - but not universal. They are experts in doing certain duties which they have been trained to do.
AGI, in its turn, could afford to:
- Acquire new skills without having to be retrained afresh.
- Move the information across areas.
- Solve unfamiliar problems
- Understand context deeply
- Make independent decisions
Concisely, AGI would not only do things but it would comprehend them.
What Would AGI Look Like?
If AGI were achieved, it might:
- Write a novel in the morning
- Afternoon. Design a scientific experiment.
- Study a new computer program in the evening.
- Planning business choices at night.
And no need of separate models, per task.
That is what makes AGI radically different to the existing systems.
Is AGI Already Here?
No, not yet.
Although the current AI is extremely sophisticated, it nevertheless:
- Lacks true self-awareness
- Lacks no goals of his own.
- Extremely dependent on training information.
- Difficult to reason to patterns not learnt.
Scientists and institutions worldwide are working on the direction towards AGI yet it is not a reality but only a dream.
Why AGI Matters
AGI could potentially:
- Increase speed in scientific breakthroughs.
- Transform education
- Revolutionize healthcare
- Reinvent work and innovation.
It is also a cause of serious concerns regarding the safety, ethics and control.
The debate of AGI is not merely technical, but philosophical.
Simple Terms
Consider the present-day AI as highly efficient experts.
AGI would be a flexible generalist, would be able to learn and think like a human being, but would be able to do so at a scale and speed potentially many times greater.
AGI will become a reality within the next 10 years or not, but one thing is definite:
It is among the most tremendous technological projects in human history.
