What Is Strong AI?

While it's still theoretical, Strong AI will be able to do anything humans can do

Strong AI is a theoretical type of artificial intelligence (AI) capable of performing any task a human can do. Although strong AI isn't a reality, now is the time to consider its benefits and drawbacks.

What Is the Definition of Strong AI?

Also known as artificial general intelligence (AGI), strong AI can communicate and reason like humans. When combined with robotics, strong AI can perform physical tasks as well. Examples of strong AI so far only exist in sci-fi movies such as 2001: A Space Odyssey, WALL-E, and Her.

Strong AI can learn new skills it wasn't initially programmed for and develop goals independently. Strong AI could presumably identify human emotions and motivations, but it's unclear whether AI can actually feel emotions. This question is increasingly important to ethicists since AI is becoming more human-like.

Two concepts closely related to strong AI are deep learning and neural networks. Neural networks comprise layers, or nodes, that perform different tasks simultaneously. Deep learning is the process of a neural network independently learning and making its own decisions. These ideas are likely key to the development of strong AI.

Strong AI vs Weak AI

The main difference between weak and strong AI is that weak AI is confined to performing specific tasks. All current uses of AI are considered weak AI.

The definition of artificial intelligence has shifted over the years. The Turing test used to be the standard for determining whether or not a computer program possesses intelligence. Developed by Alan Turing in 1950, the test involves a human subject asking another human and a computer program the same questions. If the subject can't tell which responses came from the computer, then the computer is believed to have intelligence.

While chatbots like ChatGPT and Bing AI could pass a Turing test, they are still considered weak AI today because they are limited to one task (responding to text inputs). Virtual assistants like Alexa are weak AI because they can only respond to voice commands.

What Can Strong Artificial Intelligence Do?

Theoretically, strong AI could perform any human job, from cleaning to coding. There aren't any real examples of strong AI, but the concept could have a revolutionary impact in several fields, including:

  • Entertainment: Creating music, visual art, and movies
  • Healthcare: Medical imaging, drug research, and surgery
  • Robotics: Cooking, cleaning, and childcare
  • Engineering: Programming, building, and architecture
  • Security: Detecting fraud, preventing security breaches, and public safety
  • Transportation: Self-driving cars, trains, and planes

Advantages and Disadvantages of Strong AI

Many AI engineers, including the "godfather" of AI, Dr. Geoffrey Hinton, have voiced concerns about rapid developments in the field. Aside from the potential economic impacts of replacing most human jobs, strong AI might harm humans. If strong AI can do anything a human can, then it stands to reason that it can lie, deceive, and destroy.

On the other hand, strong AI could vastly improve people's lives, freeing them from work to spend more time with their friends and family. Since AI can perform almost any task faster and more efficiently than a human, it could result in better healthcare, safer roads, and more innovative inventions.

FAQ
  • Will Strong AI take my job?

    Maybe, one day. Of course it depends on what your job is, but it's not going to happen in the foreseeable future. For now, we can take advantage of similar tools to help us do our current jobs better (think grammar checkers, code creation, photo enhancements, etc.).

  • Does my smartphone use AI?

    These smart assistants are often referred to as AI, but they likely aren't the type of AI most of us think about. You know, the kind that you have a full conversation with. And when you open your photo library and search for ladder or tree, it might seem like AI is showing all the pictures containing ladders or trees, but really that's just machine learning.

Was this page helpful?