Does AI Sometimes Mean Artificial Illusion?
The thing is, opinions vary wildly on where AI fits into the curve. For many, it is firmly in the inflated expectations phase, with almost every new software start-up saying they use AI to achieve some particular advance or new insight. Others suggest that we are already starting up the slope of enlightenment, citing applications such as conversation bots, unbeatable gaming players or autonomous driving as examples of practical, real world installations.
So who’s right? The problem is that all these views are correct because AI is not one technology but a collective term that encompasses a range of techniques and approaches. The definition of AI is as broad as saying it is an area of computer science that emphasises the creation of intelligent machines that work and react like humans.
However, with such a wide and hence fuzzy definition, there are problems. The scope of AI is disputed as machines become increasingly capable and tasks considered as requiring "intelligence" are often removed from the definition. Hence some solutions, such as character recognition, are now removed from AI as they have become ubiquitous and we now don’t perceive them as requiring any intelligence at all.
The problems are further compounded by the fact that some approaches that do fall within AI are often (sometimes deliberately) confused with it. For example machine learning is a part of AI that uses statistical techniques to give computers the ability to "learn" with data, without being explicitly programmed. Using our example of autonomous driving, the makers of these vehicles claim that they are using machine learning (and hence AI). In reality, they are actually collecting vast data sets from miles and miles of testing and using these (analytics, not really a part of AI) to find the best historic match to the real time inputs.
And this is the problem for investors thinking about AI. Much of what is out there is still very much based on how much previous data has been gathered. This is why Google has a much better chance of editing unwanted content than Facebook, simply because Facebook as only recently woken to the content issue, has little data and is consequently hiring many humans to act as editors. Hence (partly) Facebooks’ recent share dive.
Another definition of Ai is “smart” machines. However, as an investor it is always better to make sure the execs and team are smart first, as almost every piece of analysis shows that it’s the people in the enterprise that make the difference.
If you’ve liked this commentary why not link to it and see further articles