By Editor Morten B. Reitoft 

Artificial Intelligence or AI find ways into more and more services, and though we all use AI today, it might be interesting to share what it is, the obstacles, and where it will take us in the future.

Starting with where AI will take us in the future is, of course, an impossible prediction. However, the European Union (EU) has just proposed a set of initiatives to control the development of AI politically. You can read about that initiative here. The intentions of having ethical considerations and other measures might be good, but I am not sure the EU can ever enforce whatever set of rules set. The European Unions initiative is maybe also more a question of the competitive situation you see between the US and EU. Google, Facebook, Amazon, Apple, and others are all American based companies with a global reach, and with vast resources. So there is a politically, economically, and an ethically perspective to AI - of course.

There might, however, be something way more important things to consider. According to the late Professor Stephen Hawkins and CEO Elon Musk, the entire humanity is on stake. Both, of course, also see the advantages of AI. Still, if we see AI used in ways that end up being uncontrollable, we need to consider how to manage the development.

But what is Artificial Intelligence? Artificial Intelligence is not really intelligent in a human sense. AI is also referred to as 'machine intelligence." Machine Intelligence may not say much more, but is a machine's ability to learn, and therefore becoming better and better to do whatever it has been taught. One of the fantastic showcases of AI is IBM's Watson. Today Watson is a real-life example of how valuable AI is since it can review things faster and more efficient than a human. I.e. American hospitals are using Watson to review x-rays so that diagnoses can be made more quickly. In the printing industry, we see AI in more and more products. These AI algorithms are most likely less complicated and have less impact than the AI's Hawkins and Musk refers to. However, these are essential components when vendors like IMP, Tilia Labs, Octoboost, Ultimate Technographic and others are super-optimizing workflow automation and maybe in particular impositions.

With self-learning systems in the printing industry, no question that processes get faster and smarter, and give the printers who have invested in these technologies, and been able to implement these, a competitive edge. A question I ask myself is the ever-recurring question why optimization of primarily prepress is so important when the utilization of equipment is so low. I, of course, understand that many printing companies, fortunately, have a high utilization degree, and therefore have huge advantages using AI/workflow/automation.

One of the things that I would like to understand is, of course, also how various systems can work together. When I go to work, I use my computer constantly. The real value in my work, however, comes when I meet people, and we start collaborating. Collaboration is an essential part of developing new products, services, and ideas, and I am not really sure if hundreds of AI-enabled applications will be able to collaborate and deliver what humans would have been able to do together?

AI is for sure, an essential part of the future, and it will, for sure, open up for crazy exciting ideas, optimizations, and hidden potentials.
This, by the way, reminds of a very unpleasant discussion I was involved in last week. It was a Facebook user group with offset print operators. I wrote that "both Heidelberg and Komori are working on solutions that eventually will eliminate the need for operators." I can understand that you can question whether this can/will happen. I can also understand that you can be upset since you eventually could lose your job. But I can't understand that you chose to close your eyes and hope this will never happen, and I can't understand that I was blamed for this development since I was simply reporting, but that's life.

AI, Robots, IoT, Industry 4.0, workflow/automation, etc. are here to stay. Now it's a question how to take advantage of the technologies to the greater good for the entire industry.

---

We at INKISH by the way also use AI ourselves. When we transcript films, we use two services - one is called Sonix.ai, and the other is called Descript. Both services are fantastic. Sonix.ai is not as good as descript. Sonix is capable of transcribing multiple languages, as well as translating these into SRT files that are used for subtitling. Descript is a desktop application that is way better transcribing the films, and you can even use it to edit the film, from the transcribed text. Part of Descript is called Lyrebird (which is not publically available yet), that allows the user to analyze a voice and then being able to add words, correct sentences etc. with the "real" voice. Sounds amazing! Well - just a few insights.