Video Gamer is reader-supported. When you buy through links on our site, we may earn an affiliate commission. Prices subject to change. Learn more
OpenAI, the company behind ChatGPT, has in its charter the aim to create artificial general intelligence that “benefits all of humanity”. This sounds like a pretty good deal to me, but what exactly is an AGI?
OpenAI themselves only offer a fairly vague definition saying AGIs are “highly autonomous systems that outperform humans at most economically valuable work”. Exactly what counts at highly autonomous is not given a clear benchmark.
Essentially, an artificial general intelligence is an AI that is capable of approaching any task given to it through its own internal logic. It would be able to understand and learn in a way similar to humans, thinking creatively to complete tasks, rather than going through an established process taught to it. The concept is also sometimes described as strong AI or full AI.
This contrasts with weak AI, which is a form of digital intelligence which is designed for specific tasks. Think of Alexa or Google Assistant searching for something for you or setting an alarm. They are able to understand you and complete tasks but both rely on human interaction. They are intelligent in some senses but their abilities are limited to a specific field: ask Siri to catch a fish and it has no means of even approaching the problem.
Another way one could define it is that a full AI would experience consciousness though this is a whole other can of worms that we’ll leave to the philosophy majors.
In truth there is no one answer, different academics have different exact definitions and bars but it is generally agreed that it has not yet been achieved, though the GPT-4 model has perhaps come the closest.
Jasper AI
Copy AI
Originality AI detector
Are any chatbots AGIs?
The power of chatbots has taken a lot of people by surprise. ChatGPT can code, pass the bar exam and is only getting stronger.
The latest GPT-4 model, is quite possibly the most powerful artificial intelligence created so far. It has passed the legendary Turing test and has been shown to use logic to complete tasks. For example, being unable to complete a captcha test, it paid someone online to complete the test for it.
A recent research paper from Microsoft was even titled “Sparks of Artificial General Intelligence: Early experiments with GPT-4”. However, applying the term AGI to GPT-4 is hotly contested, even by some of its creators who point out it still has a long way to go.
An AGI has never felt so feasible though, and the possible implications of this are truly world-changing. This is what prompted tech leaders to ask to pause giant AI experiments arguing that its impact needs to be further studied.
What’s certain is that the future of AI is bright, and progress will continue to be made, a fact that is exciting, if also slightly terrifying.
Frequently asked questions
Will OpenAI pause AI experiments because of AGI?
It is unlikely that OpenAI will fully pause their experiment, as requested in the open letter, though we don’t know for sure. The company has acknowledged that there might be a need for an ‘independent review’ process for future AI creation, to ensure that safety concerns are met.
What is an example of AGI?
No AGIs exist in reality, however they have been depicted in fiction. Intelligent robots that feature in the likes of Westworld or Star Wars, for example, would qualify as AGIs.