There's a definite feast or famine vibe regarding how the construction industry views the possibilities of artificial intelligence (AI) in today's world, experts said at a recent webinar, The Power of AI in Construction is Real (Part Two), sponsored by Ineight. 

Potential users harbor several misconceptions about AI that hinder adoption and efficiency, panelists told attendees. The panelists included Ajoy Bhattacharya, Sr. Technologist, Microsoft; Ali Khaloo, Ph.D., CEO, Aren, infrastructure management software; Professor Eduard Hovy, Executive Director, Melbourne Connect, facilitating cooperation and innovation between academia, industry and government.

Over/Under Betting on AI

There’s been a pattern of “overestimating and underestimating” AI’s potential capabilities for many years, said Ali Khaloo, Ph.D., CEO of Aren, a company that develops software infrastructure management to handle heavy construction projects leveraging AI. 

Noting AI has been around since the 1950s, Ajoy Bhattacharya, Sr. Technologist, Microsoft, with prior industry experience bringing technology to construction projects, says he's concerned when potential users ask if AI can tell them "What's going to happen on the job tomorrow." AI doesn't work that way (at least yet), and counting on that predictive power is a red flag of concern for Bhattacharya. 

It is “just as important to know what AI can’t do,” explained Bhattacharya. It’s a “big misconception” to think AI is predictive in that manner. Such expectation “tells me people need to better understand [AI],” he says. 

AI has been strapped on a roller coaster of hype for decades, says Professor Eduard Hovy, Executive Director of Melbourne Connect. He’s been working with academics and companies, helping them to better understand and leverage AI. 

Referring to an “AI winter in the 1990s,” Hovy recalls when the hype backfired on AI and disenchanted many potential users. "There's some concern in the industry today" that potential users will again expect too much and walk away in frustration. "We are in another hype cycle now, and [the AI industry] has some concern" that users will again become frustrated because AI “doesn’t do all the magic you read about.”

Nevertheless, AI is for real and can help construction projects today in valuable areas such as cost predictions and process flow optimization, Hovy said.

"AI is here; it can be used today," agreed Ali. "We are still in a learning curve, but AI can be used today; it's not super futuristic." He noted that some leaders are scrambling (and “freaking out”) as they try to find ways to use AI now. “Some people thought it would be ten years from now, but it’s [happening] today,” he added.

 Aligning expectations with reality is a key to advancing AI, Hovy said. Otherwise, “There’s a danger there will be too much excitement followed by disappointment,” or people will be discouraged and not explore AI at all.