Third Installment: AI tomorrow
/Artificial Intelligence in the visual infrastructure-inspection industry
Advanced algorithms taking over the world. Artificial lovers in the form of sentient algorithms. When we think about Artificial Intelligence in the future, our minds often wander along the lines of science fiction.
Yet there are many other, equally exciting, prospects for the AI of tomorrow. This piece will explore some of them.
AIII_ Artificially Intelligent infrastructure inspections
Following up on the topic of the first two articles of this series, we will look at AI in visual infrastructure inspections. Getting a glimpse of the future applications of AI in the visual inspection industry, it is as simple as browsing websites or watching a webinar. Companies are eagerly sharing their visions for this technology, and we are expecting those visions to become reality within the next 5 years. So, let’s break down what those visions are. In 5 years, you can expect an AI-algorithm to answer the following questions:
Where is my asset? Determining the exact geo-location of an asset is already possible and practiced. It will likely become a stable part of any data management software in the next years.
What is around my asset? Mostly, the answer will be vegetation. There are several ways of managing encroachment today which include AI, like LiDAR and Photogrammetry, and it is likely that we will see even more technologies emerge in the coming years. The existing solutions will also develop to rely more on AI, than they do today, for both data capture and processing.
What needs maintenance? To determine the condition of an asset, algorithms need to get a lot better at identifying faults and their severity. We covered the challenges of this subject in depth in our previous article; AI today.
How should I prioritize resources? This is a complex question. It requires both accurate asset condition scoring, and predictions based on trends and historic data. Whether an algorithm will be able to perform this task well within 5 years, is hard to say. What we can say with certainty, is that a lot of companies are working towards this goal and the industry will get there in the foreseeable future.
Most of these applications are already available in one form or another. What will happen in time, is that they will grow more reliable, advanced, and become more widely available. All the capabilities above, will likely be part of the same data management platform. This makes sense, as gathering all your data in one place has advantages when training AI-algorithms.
Now what happens if we look beyond the platform and further into the future?
Though anything is theoretically possible, there are scenarios which are more likely to become reality in the next 10 years. Those are the ones we will focus on.
AI is all about finding a way to perform today’s tasks faster, and with better results. Below are some setups, we find likely, that AI will enable.
Automated inspections
As drones are getting smarter and BVLOS regulation is moving forwards, fully automated inspections are almost inevitable.
Fleets of cooperating drones will be dispatched at pre-determined intervals, to perform fast and accurate inspections of infrastructure. Image-stabilizers will be AI-powered and ensure the highest data quality, and object recognition will be live, so only faulty equipment is documented. Drones will automatically return from missions and upload data to the company platform, for further processing.
Once this becomes reality, the only real bottleneck will be maintenance. This will likely be solved by adding AI-powered repair-drones to the maintenance crew.
At this point, where inspections and maintenance are both fully automatic, the return on investment for utilities will be immense.
Digital assistants
You won’t see the ROI from automated inspections repeated here. Digital assistant will be born, to make your work-life easier.
Imagine Siri’s smarter grandchild. She will make user interfaces, as we know them, obsolete. Just ask your assistant a question about your assets and let her present the answer to you, in the most logical way.
Deeper algorithms have already sorted, analysed and prioritized the data. The assistant has all the knowledge of trends, historic data, and weather, that is key in assessing risk. This will enable her to make better predictions, instantly.
Just imagine the ease of managing governance, risk, and compliance with a tool like that.
AI setting the direction
At some point, algorithms won’t be trained by humans anymore. At some point, new algorithms will be birthed and raised by other algorithms.
The systems that we set in place, will know so much about our operations, that they will begin to detect trends we have missed. It will recognize more efficient ways to obtain our desired results. It will identify missing links and fill in the gaps on its own.
In short, Artificial Intelligence will stop assisting in-, and start driving development.
And then, what about us?
What will happen to the concept of “jobs” as we know it today? This is a question, which preoccupies many minds and lead the resistance against AI.
The answer to this, involves even more guesswork, than the paragraphs above. What is safe to say, is that we will have to re-think the way we earn a living today.
Some people speculate in more leisure time, and believe that when we do work, we will do so primarily out of interest.
Others think new types of jobs will emerge. Positions, that we don’t have the imagination to think up today.
Or maybe we will become power-plants, existing only in a virtual reality.
What do you think?
Article written by: Scopito
