The Atlantic published an article titled ‘Artificial Intelligence’ Has Become Meaningless. They are not saying AI is actually meaningless, but rather touting a software solution as being AI in so many cases does not actually mean the application uses real AI. As they put it, “…in most cases, the systems making claims to artificial intelligence…[are] just software.”
To a large extent, I blame the marketers. Marketers use popular terminology to describe their products, whether they really apply or not. This happens with technology products targeted for high end purposes, not just consumer products.
In the literature discovery field, there was a time when “semantics” was the hot term. Semantic search, semantic web, semantic ontologies, and so much more. But, what did this actually mean? Applications were being touted as being “semantic” whether they really were or not.
Today, it is the phrase “artificial intelligence” that is being misused. And the points made in The Atlantic article are spot on stating that AI is “often just a fancy name for a computer program”.
How can you tell if the claim of an application being AI is valid, or just unabashed marketing? Here are some questions you can ask:
- What AI algorithms are being applied?
- Does the system continue to learn?
- Is critical information relevant to the user’s need automatically defined from the content?
And, just because we know you will ask: Yes, Qinsight indeed uses neural networks and several other AI algorithms, Qinsight continues to learn, and Qinsight extracts information to provide you the most relevant documents on-the-fly.
You can also spot the real AI folks by saying you have “haphazardly run across deer in random forests”. If they give you a blank stare, they don’t have a clue about AI.