What is the impact of AI on the world of inspections?

Artificial Intelligence has been on the rise over the past decade. Its applications range from your phone pointing you to the best restaurants nearby, to beating humans in chess and cars that drive themselves. The AI that we know and use in our daily lives is called Artificial Narrow Intelligence (ANI)1. ANI is very good at doing one specific thing and continuously getting better at that: all of the applications mentioned above are examples of ANI.

 

Currently, the TIC (Testing, Inspection and Certification) industry is largely reliant on skilled workers that perform manual inspections. Health, safety and environmental inspections require thorough power of observation as well as extensive knowledge of procedures, machines, tools and possible defects and their dangers; a lot of expertise is required to perform quality inspections. ANI can assist in getting inspections done safer, faster and more accurate. Down below, we’ll show you a few ways in which ANI can help inspection companies create a more safe, secure world for everyone.

"ANI can assist in getting inspections done safer, faster and more accurate"

Computer Vision & AI

Computer vision focuses on capturing and analysing pictures, and has been around for decades. Since the manufacturing price of optical lenses has dropped significantly in recent years, computer vision technology is now relatively affordable. As you might imagine, ANI and computer vision go hand in hand. This is due to the fact image and pattern recognition are a big focus for those studying artificial intelligence. Consequentially, most AI improvements are made in the area of visual detection.

 

There are several benefits to combining AI and computer vision. First of all, artificial intelligence improves upon its own detection capabilities, becoming more and more accurate in recognizing dangers and defects. Second, computers and lenses don’t get tired! Everyone has an off-day every now and again. A foggy brain, eye strain, distractions; they happen to the best of us. Machines don’t suffer such conditions (assuming they are kept in good shape).

 

This means that anything detected by a camera coupled with an algorithm is likely to be much more accurate than what the human eye can see. The concept of computer vision and ANI can be applied in many different ways. For example, many factories already use it to automate the quality control of their manufacturing process. This specific application of computer vision and AI is called ‘Machine Vision’.

 

The big benefit for the TIC industry seems to be in the application of computer vision in drones. Think about it: instead of having to climb up to or down into a small or dangerous space, you simply send in the drone to take pictures and analyse them in real-time on your tablet (or other mobile device). You’ll have all the information you need during your inspection; the AI can give you precise measurements and analysis to help you determine the next steps.

"The combination of Machine Vision and an algorithm is likely to perform more detailed inspections"

Planning & Scheduling

Planning and scheduling can be a very complex business. The more people, materials and actions you need to take into account, the more daunting it gets. Artificial intelligence has the potential to make this a lot easier. As I’ve stated earlier, pattern recognition is something that AI is really, really good at2. Considering this, it makes sense to apply ANI to work planning. It can recognize and calculate an inspector’s most optimal route for the day, plan the right tools along with it and -combined with report data- automatically plan follow-up actions in the blink of an eye.

 

Applying ANI to planning has the potential to save planners headaches and lots of puzzling. Additionally, it ensures a fair, balanced planning. After all, a machine that plans off of pure, factual data won’t play favourites! They simply plan and schedule in the most optimal way. All it needs is a little input: types of actions to plan, amount of people and their expertise, amount of hours available for scheduling, etc. etc. It’ll then take care of the rest.

 

Another benefit of applying artificial intelligence to work planning, is that it can automate inventory and purchasing. If set up right and given the right information, ANI can assist in keeping track of stock levels of materials and tools and automatically order stock replenishments and repairs. At the risk of it sounding almost too good to be true: if you combine this with the aforementioned capabilities for workforce scheduling, you’ll never have trouble planning a job or project ever again!

"Applying ANI to planning has the potential to save planners headaches and lots of puzzling"

Analytics

More data than ever is available nowadays. Far more than a single person could ever meaningfully use to perform their work or make decisions. AI is a powerful tool in compiling relevant data and filtering, arranging and presenting that data in a visually clear interface. In short: AI can quickly give you the exact data you need to make informed decisions.

 

As noted before, computers are far better and identifying patterns and trends than humans are. Imagine if I gave you a 1000 pages of inspection statistics and told you to ‘make a ranking of the most common deficiencies over the past 10 years. I imagine I probably would not get invited to your next party. Artificial Intelligence has no trouble going through all the information, seeing the patterns, organizing the data and even, based on previous calculations, giving advice on what to do with the information.

 

Going a step further and allowing artificial intelligence a degree of decision-making, a computer could actually use datasets to determine what to inspect when, and why. This is especially useful in the management and planning departments.

 

These are just some simple examples of applying artificial intelligence to (historical) data sets. The bigger point is: no matter what you want to know, an AI is capable of giving you all the relevant data you need at the click of a button. The only limits are the questions you ask it and the information you feed it!

"A computer could actually use datasets to determine what to inspect when, and why"

As the technology matures faster and faster, we’re looking to implement the capabilities of Artificial Intelligence in our own inspection software over the next few years. If you want to know more, you can download the brochure here.

1 Artificial General Intelligence (AGI) -an AI attaining a level of understanding the same as the human brain- is the next step in the development of artificial intelligence. When AGI becomes real, it will likely cause a societal shift unlike anything we have seen so far. If you have some time and you’re interested to learn more, I highly recommend this article by Tim Urban. It’s a long read and couple of years old, but its main points are as relevant today as when the article first appeared in 2015.
2 A lot better than us humans, who often think we are very good at this. More often than not, our pattern recognition and subsequent decision-making is based in biases. Computers have no such problems, providing they are not designed with inherent biases.
en_USEnglish