Everyone is talking about AI these days, especially in the wake of COVID-19—we’re all looking towards automation to make our processes safer and more efficient. Manufacturing is the perfect field for AI to make significant differences, because AI is great at perfecting and improving repeatable processes, from assembling products to ensuring consistent quality.
The recycling sector of the manufacturing industry has been overlooked more recently. Although we’ve seen a massive increase in the usage of eco-conscious materials, reusable versions of usually disposable products, and simple awareness of the trash problem in the developed world, recycling has not received as much positive press lately as it did in the 1990s and 2000s. One reason for this turnaround is that massive amounts of recyclable materials are not actually recycled, even when they are sent through municipal recycling systems, and this is often because it can’t be sorted accurately and efficiently. Both the plastics manufacturing industry and the recycling industry are experiencing lots of pushback due to their environmental impact: plastics are perceived as having too great a negative impact, and recycling as having too little a positive one.
But recyclables, especially plastics, aren’t going anywhere, especially in the face of a global pandemic that necessitates items like gloves, masks, and disposable bags. No one is willing to forgo their personal safety for the sake of avoiding plastic, so we need a way to recycle more items with better accuracy and speed. Both plastics and recycling can help improve their images and make a positive impact on the environment simply by utilizing AI as part of their sorting processes.
AI algorithms have become great at detecting specific objects in noisy images and videos, and at auto-labeling those images and videos based on predefined taxonomies. The best place for plastics engineers and plant managers to start applying AI and getting quick returns is by using these object detection technologies to sort recyclables into different categories automatically, or to identify product defects early on using image recognition models to minimize production waste and to automate performance monitoring of plants by creating digital twins of production processes and equipment. Both segments of the manufacturing industry can achieve lighter environmental impact while simultaneously saving on costs by reducing waste, all by applying the same types of solutions. Of all these possibilities, the most impactful would be in leveraging AI to sort products for recycling. Current processes are inefficient and error-prone, and can even present health risks if they rely on humans to participate in close contact with machinery or waste products.
We all need plastics. There’s no denying that, especially in a post-pandemic world. We also know that plastics can be bad for the environment, and we know that no matter how hard we try, human habits are difficult to change. Plastics are here to stay. With that in mind, the foremost objective of the plastics industry needs to be to minimize the impact of production on the environment and to recycle as much of their product as possible with reliable, efficient, and safe recycling systems that can be easily replicated in manufacturing plants across the globe. By implementing AI to do these tasks, plastics manufacturers could reduce waste, recycling plants could process far more materials, and both industries could improve their standing in the public eye.
