AI Still Not As Smart As Humans According to This Expert

The jury is still out on whether or not new innovations like ChatGPT are going to be a net positive or a net negative for society with all things having been considered and taken into account. The advancements of this tech are hard to deny, with many saying that AI will soon surpass human beings in terms of intelligence. In spite of the fact that this is the case, an expert by the name of Anthony Chemero seems to think that AI still has a long way to go.

As a professor of philosophy and psychology at the University of Cincinnati, Chemero has unique insights into the manner in which human cognition seems to function. With all of that having been said and now out of the way, it is important to note that many people seem to conflate linguistic ability with overall intelligence.

AI has the ability to lie and deceive people. Many see this as intelligence because of the fact that this is the sort of thing that could potentially end up passing as a facsimile of cognition, but Chemero is of the opinion that AI simply can’t reach the level of intellect that humans naturally seem to possess. In a paper he published in Nature Human Behavior, the professor argues that because AI is trained on datasets, it inculcates many of the biases present within the humans that created that data in the first place.
Large Language Models are able to mimic human speech patterns, thereby convincing quite a few people that it possesses legitimate intelligence. However, AI utilizing LLMs isn’t technically aware of what it’s saying. It simply stitches grammar and syntax together based on the data it has gleaned, and it requires much more training than humans to maintain a passing semblance of our intellectual capabilities.

Just take hallucinations for example, which is what people in the industry call it when AI spits out information that is patently false. The AI is simply calculating what the most probable next word in a phrase is, without really understanding the true meaning behind these statements.

Of course, AI is still in an extremely nascent stage of its developmental journey. As a result, it may be able to overcome these intellectual hurdles and achieve true intelligence, which itself comes with its own fair share of problems.

Read next: From Fear to Preparedness A Deep Dive into Consumer Attitudes Towards Cyber Threats and AI Misuse
Previous Post Next Post