Bringing Artificial Intelligence to Manufacturing

Element AI SVP talks about bringing AI to the rest of the world.

Talk of automation in manufacturing tends to focus on industrial robots, for obvious reasons.

The robotics market is growing at an unprecedented pace and vague worries about job loss due to automation—however misguided—often take shape in visions of robots replacing individual workers on production lines.

However, there’s a much less tangible form of automation that’s poised to make an even bigger impact on manufacturing in the near future: artificial intelligence (AI).

The concept is notoriously difficult to pin down. “I would say it’s as difficult to define as intelligence itself,” noted Philippe Beaudoin, SVP research at Element AI. But for manufacturers, what matters most is what AI can do.

The technology is already seeing applications in construction and additive manufacturing, as well as self-driving vehicles and industrial robotics. Broader applications for AI, ones that could be particularly useful in manufacturing, include analyzing large datasets and predictive maintenance.

Of course, most manufacturers aren’t interested in becoming AI experts, which is where Element AI comes in. Offering “AI as a Service”, the company provides applications that allow users to leverage artificial intelligence to tackle a variety of challenges, from speech recognition to automated decision making.

Element AI's Mur.AI takes a reference image and processes a video stream in real time to give it the same style. In this case, the reference image was this mural, painted by Montreal-based artists A’shop. (Image courtesy of Element AI/the author.)

Element AI’s Mur.AI takes a reference image and processes a video stream in real time to give it the same style. In this case, the reference image was this mural, painted by Montreal-based artists A’shop. (Image courtesy of Element AI/the author.)

Element AI primarily works with clients that already have large data sets—such as the shipping schedules for busy ports or production data from factories—rather than gathering that data independently. This lets the company focus on building AI that can, for example, interpret high-frequency time series from distributed sensors on a production line to make maintenance recommendations.

As an added benefit, this allows Element AI to improve its systems with each new project the company takes on. That approach seems to be paying off: almost a year ago, the Montreal-based start up had eight employees. Today, after over $100 million USD in series A funding from the likes of Intel, Microsoft and NVIDIA, Element AI has a staff of 160.

There’s clearly a lot of enthusiasm here, but one might worry about the impact something as extensive as AI could have on manufacturing jobs. Beaudoin emphasized the cooperative role employees can play, particularly when it comes to potential bias in the data sets his company is using.

“Having a human in the loop is really important,” he said. “You can’t hide behind the fact that it’s an algorithm, because the machine is just capturing patterns in the data. So, you need to know your data.”

This raises questions regarding the ethical implications of making decisions based on machine learning. We’re not talking about hackneyed, Terminator-type scenarios, but more realistic concerns about, for example, the potential discriminatory aspects of automated decision-making. There’s even an industry organization that was created to deal with these issues.

For his part, Beaudoin expressed concern about this problem, but was careful to point out that its scope extends well beyond his own company. “The ethical problems are real, but they’d be out there with or without Element AI,” he said.

For more information, visit the Element AI website.

Written by

Ian Wright

Ian is a senior editor at engineering.com, covering additive manufacturing and 3D printing, artificial intelligence, and advanced manufacturing. Ian holds bachelors and masters degrees in philosophy from McMaster University and spent six years pursuing a doctoral degree at York University before withdrawing in good standing.