Strong artificial intelligence (AI) is a theoretical form of AI that describes a particular AI development mindset. Strong AI requires machines to be as intelligent as humans. They should thus be self-aware or able to independently solve problems, learn, and plan.
Strong AI is also known as “artificial general intelligence (AGI)” or “general AI.”
Read More about “Strong AI“
Based on the short definition above, strong AI hasn’t turned into reality. Even the best and brightest AI developers have yet to create a machine indistinguishable from humans.
What Is the Goal of Strong AI?
Strong AI aims to create a machine that would have to learn through inputs and experiences. It should thus constantly progress and advance its abilities over time.
What Tests Should a Machine Pass to Be Considered a Strong AI System?
Experts believe a machine can only be considered a strong AI system if it passes the Turing Test and the Chinese Room Argument (CRA).
Turing Test
Developed by Alan Turing in 1950, this test compares how a machine behaves to a human. Dubbed the “Imitation Game,” an “interrogator” will identify differences between computer- and human-generated outputs by answering a series of questions. If the interrogator can’t definitively tell which result was given by the machine and which by the human, the device passes the test. If not, the machine can’t be categorized as intelligent.
A strong AI system needs to perform various tasks equally well. As such, the Turing Test should evaluate its textual, visual, and auditory performance compared to a human.
CRA
Created by John Searle in 1980, CRA states that computers can’t understand and think like humans. In this test, the machine is tested against a human who doesn’t speak Chinese. The person will be made to sit in a closed room where a book with Chinese language rules, phrases, and instructions is placed. Another person fluent in Chinese passes notes written in Chinese into the room. Using the language phrasebook, the person inside the room should choose the appropriate response and pass it back to the Chinese speaker.
While the person in the room may be able to provide the correct response, that doesn’t mean he or she understands Chinese. Thus, he or she will only be considered “strong” when he or she starts responding without the phrasebook’s help.
In parallel, a strong AI system should have consciousness or understanding, not just respond to the person in the room by translating words.
What Is Weak AI and How Does It Differ from Strong AI?
Weak AI or narrow AI systems can only perform a specific task. An example would be a computer that can play chess as opposed to strong AI systems that are capable of multiple functions and can eventually learn to solve new problems.
Weak AI systems need human inputs to learn. Strong AI systems don’t necessarily need human inputs though these can speed up their learning. Self-driving cars and virtual assistants are also examples of weak AI systems.
What Characteristics Should a Strong AI System Have?
A machine can only be considered a strong AI system if it has the following human characteristics.
- Can generalize knowledge and apply it to different circumstances
- Can use knowledge and experiences to make plans
- Can change and adapt to different circumstances
- Can reason out
- Can solve puzzles
- Can exhibit common sense
- Has consciousness
- Can solve problems beyond mathematical equations
- Can discern needs and emotions
The following image shows what a strong AI system should be able to do.
—
As mentioned earlier, we have yet to see a strong AI system—a computer that can’t be distinguished from a human. While some believe we’ll witness its creation soon, most think creating such an intelligent machine will never happen. So much for robots taking over, right?
Key Takeaways
- A strong AI system is a machine that can’t be distinguished from a human.
- To be considered a strong AI system, a machine should pass at least two tests—the Turing Test and the CRA.
- Weak AI systems can only perform a specific task, while strong AI systems can do several jobs.
- A strong AI system should be able to generalize knowledge and apply it to different circumstances, use knowledge and experiences to make plans, change and adapt to various circumstances, reason out, solve puzzles, exhibit common sense, be self-aware, solve problems beyond mathematical equations, and discern needs and emotions.
- Strong AI remains a theory to this day.