AI in Manufacturing

Artificial Intelligence in Manufacturing

Artificial Intelligence. If you are reading this, it is probably for one of two reasons: (1) you find AI fascinating and read or watch everything that comes your way, (2) you work in manufacturing and believe that AI can be useful for your business.

If you are in case 1, it might seem that the use of AI in manufacturing is not very interesting, at least when we compare it with the brilliant results of AI learning to play online games and defeating all human champions, driving a car autonomously or help understand how proteins fold. Manufacturing is still widely associated with greasy, heavy, old-looking machines — not very smart at all — and repetitive processes that don’t need any fancy AI to work.

In fact, in the very realistic and down-to-earth manufacturing realm, the AI can lose some of its lustre and let it show how, internally, it’s not much more than statistics and mathematics. There is a saying about AI applications: it is only true AI if it is not done yet. Chess was, not that long ago, the board where the most state-of-the-art AI algorithms played. But who does think nowadays that chess playing is real AI? As Rodney Brooks once said: “Every time we figure out a piece of it, it stops being magical; we say, ‘Oh, that’s just a computation.'”

In reality, there are many fascinating topics surrounding the use of AI in manufacturing. It is a real and very important environment to begin with, and the application of AI has a lot at stake, for example helping to reduce the carbon footprint and thus helping to avoid a climate disaster (and manufacturing has a lot to improve, see Figure 1 below).

Figure 1: world CO2 emissions per sector, the others are buildings (7%), transportation (16%), agriculture (19%) and electricity (27%). Source: https://www.breakthroughenergy.org/

If you are in case 2 above (interested to know more about the applications of AI in manufacturing), you should be aware that there is an overuse of the term Artificial Intelligence; it is now one of those keywords that is used everywhere. This has not always been the case, in the last decade of the 20th century, very few people still considered AI as something with a bright future. This was due to the exaggeration of the achievements that we researchers and the media made (and this is one of the recurring problems with research in AI).

On the contrary, since 2013 AI is again modern and ground-breaking and, once more promises to change the world. This is the year when Deep Learning started to attract public attention with achievements that seemed to be a long time in the future (like autonomous driving, image recognition, speech recognition etc). But where are, exactly, those opportunities of AI in manufacturing? Are they really going to change the industry much? What are the most prominent use cases of AI in the sector? In this blog we will try to answer those and other questions, always keeping a stance as balanced as possible, without any hype to help manufacturing companies get a clear-glass view of AI applications.

We will present how AI is actually used in day-to-day tasks, talk about its successes but also of its failures. Because applying AI and achieving good results is not easy. According to the 2020 report “AI in European Manufacturing Industries”, commissioned by Microsoft and conducted by EY, only 12% of surveyed companies have managed to successfully scale AI company-wide, and report having achieved significant internal and external value from implementing AI, but only after investing in AI for more than 5 years, on average; companies that have been working with AI for less than 12-18 months have yet to see significant results. This clearly shows how AI is not a magical ingredient that you simply add to your crafting recipe. It takes time, good data, good people, and full support at all levels of the business to achieve results. The good news is that once you are in a position to reap the benefits, these can have a profound impact.

This article is the first in a series of articles that will present the basics of AI and the current and future applications in manufacturing. Among those, we will touch ideas like:

• Learning from your data
• AI and humans
• AI and robots
• Intelligent control
• Sustainability

Welcome to the IMR AI in manufacturing blog, I hope you will find at least a few interesting ideas here over the coming months.

See how you can deliver impact from Industry 4.0

or

Contact Us