Defining Artificial Intelligence
To truly fit the definition, AI requires machine learning; the program must be able to seek and find information that it didn’t already know. However, most people think of AI as something interactive, and that doesn’t necessarily demonstrate machine learning.
As a result of this disparity, it’s easy to describe as AI any interactive device, product, or service, and consumers willingly and routinely accept these descriptions. There are products considered AI that may not be the real deal.
Since AI has become mainstream and demand exists, many companies are working feverishly to capitalize on it. In some cases, they take advantage of the popularity of AI and consumers’ confusion and misunderstanding about the technology to offer something enticing while not authentically AI.
This faux AI falls into the following categories:
- “Powered By” AI. More and more devices come with a voice-recognition component. From smartphones to home security, cars and bikes, everything is being linked to a voice-assist system. The products are accepted as AI, but they are not. While the system is AI, the device is only riding piggy-back on that system; the product is not AI but powered by the technology. These devices are programmed to reference AI software; they’re AI hitch-hikers.
- Just Big Fakers. Another category of not-quite-AI services is more nefarious. Believe it or not, some of the chatbots that are lauded as the most common AI around are powered the old-fashion way; by human beings. There are many instances of companies, particularly startup online companies, boasting 24/7 online chat that found implementing AI too difficult or expensive. Instead, in these instances, it is cheaper, at least in the short run, to hire people to type back pre-planned responses to the most common of questions.
- False Advertising. Finally, there are programs out there that claim to be AI that just, well, aren’t. They may have a vast array of expected inputs and tailored responses to each of these inputs, but at the end of the day, they’re just a database of inputs leading to programmed responses. These may seem like intelligent devices to 90% of users, but what they lack is the machine learning that differentiates AI. If the user inputs an unexpected or unplanned command, the device will not know what to do, will not learn what to do, and will give the same response forever to that unknown command.
As AI continues to take the world by storm, companies are scrambling to get on the AI bandwagon, and many of them do so successfully. However, like any trend, there are those that will get on board in the easiest way possible, even if that means bringing a fake ticket to ride the train. Make sure to read the fine print, think critically when advertising promotes AI products, and question whether the technology really learns or just interacts.Last modified on Monday, 11 November 2019