Weak AI (sometimes called narrow AI) is simply intelligent behaviour programmed into a machine for specific purposes. These systems are not sentient, self-aware or capable of AGI. Someone who is programmed to just do certain things within his capacities.
Examples of Weak AI
Narrow AI is used in self-driving cars to identify objects (vehicles, pedestrians etc), and interpret the rules of road. But they are no where near as capable of including a general reasoning or complex problem-solving like driving activities.
What is Strong AI?
Narrow AI: No such superhuman general intelligence capability exists now; however, a strong AI would be able to understand and learn any intellectual task that a human being can.
Sentience: In the same vein, a strong AI machine would have sentience — it could think and reason in some sense of the term.
Autonomy: Systems able to definite action without human intervention. This can range from guided systems (which provide board for back and forth interaction) that are execute a full system with no humans involved in real-time, or automated actions running somewhat independent of one needed guidance).
Differences between Weak AI and Strong AI
The main contrast between Weak AI and Strong AI is around the area of intelligence achieved. Weak AI specialized for narrow tasks, as opposed to strong AI that would apply generalized reasoning in a multiple disparate fields like humans do.
Weak AI system: A weak AI is an algorithm-driven shell computer which runs within a powerful framework so as to do all that it should have in relation. They can be taught in their realm (i.e., machine learning models) but such leaning is fixed to tasks it was trained on. In contrast, strong AI could interpret and learn in a looser more human-like way by adjusting to the novel situation without fixed algorithms.
Weak AI is completely devoid of any kind of consciousness or self-awareness. It does not comprehend its method it act, and the context in which to do so. Theoretically, strong AI would be capable of understanding it has an environment along with context and purpose.
Though Weak AI capable of accomplishing tasks without human intervention (in its domain), it is still programmed and sometimes monitored by humans. Strong AI could act on its own, think independently much like humans in many areas even without regular human intervention.
Current Status of Weak AI
In the case Strong AI were one day to be created, it would present a major problem of ethics and philosophy. Among other controversial topics debated are the rights of artificially created beings, unchecked AI autonomy and whether AI can become an existential risk. Moreover, the ramifications of misuse in war and surveillance scenarios could help determine whether or not human beings will be upended from anthropos — if such problems prove to even rise out of control.
