Buzzwords! Marketing loves them, insiders and experts hate them and yet they are used inflationary. AI, Blockchain, IoT, neural networks and deep and machine learning are hot. The only question is whether they are really part of the product - or rather part of marketing.
Just years ago, the term "digital transformation" or "digitization" was a signal that the signs of the times were recognized and that everything was being done for progress. The umbrella term was suitable for all kinds of things: Orders that were once sent by fax are now sent digitally by e-mail. Employees now discuss everything via chat instead of in meetings. But then you still want to tell the product owners exactly how and when they should do what.
AI at any price
The problem lies in the nature of the terms "digitization", "new work" or "sustainability". They are very vague and have to be used for so many things. It is often not evaluated whether it was necessary to send the orders as documents at all or whether an automated merchandise management system would not have been smarter. "If you digitize a shit process, you have a shit digital process," summarized: You can't solve your processes with buzzwords. And you shouldn't sell your products with Buzzwords either.
Nevertheless, it is done, and basically the technologies can't help turning them into buzzwords. AI, neural networks, block chain, computer vision or quantum computing are important technologies that will change our lives faster in the next ten years than in the last 100 .
But Buzzwords sell products
Being a progressive tech company today, and especially an innovative start-up, without using AI or machine learning often seems impossible from a marketing point of view. Google has confirmed its focus on AI 2018 at its in-house conference I/O, Amazon offers AI frameworks for its AWS cloud and Microsoft wants to make multiple blockchains manageable on its Azure platform.
Big Tech is setting the trend here and developing according to Willy Brandt's motto: "The best way to predict the future is to shape it. In order not to be considered old-fashioned, they want to be part of this and include "artificial intelligence" or "data analysis" in the product description or the pitchdeck for investors. There is only one shocking catch here: you have to deliver what you promise.
Algorithms are not yet AI
The development of a real AI is incredibly expensive, time-consuming and paved with pitfalls. Microsoft and others had to find this out painfully several times when their AI was fed with data during the learning process, only to be shut down again. The database was to blame here despite a lot of different information. To develop AI is very difficult, because literally you try to create an intelligence that is artificial. Everything else is not really an AI and in many cases not even close to it, even if one likes to use the term.
An algorithm that evaluates a lot of data at Amazon to indicate a probability that I will eventually buy a certain product, for example. This is not an AI (and Amazon does not sell it as such), but a complex algorithm. However, it is precisely these isolated functions that are often touted as AI by other companies.
The term has become a buzzword because it has been so watered down that its original meaning is no longer applicable: a learning intelligence that can independently give recommendations for action and develop completely new ideas. And at some point it will develop its own consciousness and become the second intelligent life form on this planet after Descartes' "I think, therefore I am".