When people talk about artificial intelligence (AI), most don’t realize there are different types of AI. Weak or narrow AI is the only one that exists today. Strong or general AI will be achieved when machines have the cognitive abilities that humans have without any human intervention. Let’s take a look at the possibilities for general AI and nine practical examples.
What is Strong (General) AI?
First, strong or general artificial intelligence is only theoretical at this point. It’s the artificial intelligence that writers have fictionalized for years in sci-fi stories such as “The Terminator” and “I-Robot.” Ultimately, when we achieve general artificial intelligence, machines will have consciousness and decision-making skills—full human cognitive abilities. The machines will not require humans to input programming to function. For all intents and purposes, this would be when machines act, feel, respond and think just like humans. We will be able to say strong AI has a mind of its own and will be able to accomplish any task it sets out to complete, just like any human. Unlike narrow artificial intelligence that classifies data and finds patterns, general AI uses clustering and association when processing data. General artificial intelligence will also be self-aware. However, just like a child, the AI will have to learn through experiences to constantly advance knowledge and skills over time.
9 Practical Examples of Strong/General AI
Simply, any task that a human can do could be accomplished by general AI. It technically has all the potential of a human brain. It could tackle any problem or task in any area, whether it be music composition or logistics—all the potential actions humans can perform. This includes the following intellectual tasks:
- Generalize knowledge and apply it as applicable to different circumstances
Humans learn from experience. They take learning from various experiences and apply it as it makes sense to other situations they encounter. This would be an example of strong AI.
- Use knowledge and experience acquired to plan for the future
Another ability humans have is to use their life experiences in order to plan for the future. As they encounter more experiences, they can use those experiences to create a plan and drive the future. In narrow AI, the machines must rely on humans to program actions. The machines are not capable of putting a plan in place for the future.
- Alter and adapt to circumstances as things shift
General AI machines will be able to adapt as they encounter circumstances. Narrow AI is only able to respond to variables that were programmed into algorithms. General AI can make decisions on the fly.
- Ability to reason
Unlike narrow AI, general AI will be able to reason. General AI machines will be able to examine a situation and determine a course of action even if it’s outside the bounds of what a human taught it.
- Solve a puzzle
Certainly, AI algorithms have competed and won video games and chess matches. Those successes are examples of AI following patterns and programs. There are some perplexing issues that don’t currently have success defined. When machines can solve a puzzle of this nature, general AI will be achieved.
- Exhibit common sense
Another very human attribute is common sense. When a machine isn’t able to rely on programming for answers, sometimes common sense is necessary. Weak AI doesn’t exhibit common sense. To be on par with human cognitive ability, machines will have to exhibit common sense.
Machines need to be capable of consciousness and be self-aware for general artificial intelligence to be achieved.
- Beyond mathematical equations
Narrow AI proved in many ways that a lot of the problems we solve as humans are just mathematical equations. When a machine can go beyond mathematical equations to general problem solving, machines will have human intelligence.
- Discern needs and emotions
General AI would also be capable of reading needs, emotions, thought processes and beliefs of other intelligent entities. This is referred to as the theory of mind level AI. There’s nothing to simulate or replicate with this type of AI, but rather machines truly understand humans.
We’ve learned a tremendous amount about the human brain, but there is still an enormous amount to understand. To create general artificial intelligence, it will be essential to comprehensively understand the human brain.
If Fujitsu-built K, one of the fasted supercomputers ever built, is any indication, it will take us years to be able to figure out general artificial intelligence. It took K 40 minutes to simulate one second of neural activity using 82,000 processors. Until general AI is achieved, there are many tasks narrow AI will continue to do for us.
Where to go from here
If you would like to know more about , check out my articles on:
Or browse the Artificial Intelligence & Machine Learning library to find the metrics that matter most to you.