Experts discuss how combining AI and IoT is creating massive opportunities, but acknowledge security problems as well

This year’s COMPUTEX Forum, themed Pervasive Intelligence, concluded on Wednesday afternoon with the AIoT Session at the Taipei International Conference Center. The final panel of the forum featured six distinguished speakers from global industry leaders discussing the latest developments in artificial intelligence (AI) and the Internet of Things (IoT).
“COMPUTEX is not just a sourcing platform, not just for procurement, it’s also for building a global technology ecosystem… Today, AI and IoT is very important,” said Walter Yeh, CEO and president of the Taiwan External Trade Development Council, organizer of the forum, before introducing the first panelist: Joe Jensen, vice president of the Internet of Things Group and general manager of Retail, Banking, Hospitality and Education at Intel.
“Over 20 years now I’ve been coming to COMPUTEX and I’ve built some tremendous partnerships and relationships with many of the leaders in this room. Our business has grown tremendously, and I think it’s grown together. We face now and will talk today [about] really doubling or more the opportunity we’ve all thought of as the future,” Jensen said.
“It’s a massive new and large and growing [field],” Jensen said. “If you go step back two or three years ago, the analysts were thinking IoT is all about some little things … billions of things generating data, spewing this data to the cloud and the analytics being done in the cloud.”
“We now realize, and the industry realizes, that IoT, especially IoT for business and for business insight and value, is really about unleashing computer vision and advanced AI technologies against large, massive sets of real-time data being generated at the edge,” he said. “These data sets are so large we’re not going to have the ability to send all that data to the cloud for analytics and processing.”
“A huge amount of those insights will be generated live real time at the edge, and that creates the opportunity for us in IoT and the edge,” he added, describing how devices,
such as sensors, will pool that data and process it at the edge, before sending the insights from that data back into industry processes and the cloud.
He highlighted three vertical markets being transformed by computer vision and AI: retail, smart cities and industrial. These markets have been increasingly employing sensors to create a massive amount of data, which are then used to improve efficiency and safety.
“There are over two million factories in China alone with no data, no analytics and really no computational infrastructure whatsoever,” Jensen said. “Think about the opportunity to go bring analytics and data to improve the efficiency and operation of those two million factories.”
“We can’t do this alone… It takes the entire ecosystem for us to bring this value to market,” he said, adding that another objective is to help customers scale faster, with market-ready solutions instead of one-off engineer products.
On the other side of the equation, those enormous opportunities can create a security nightmare.
Mike Gibson, vice president of Threat and Vulnerability Research at Trend Micro, said that the biggest issue that keeps him up at night is IoT security, particularly three challenges: The discovery of new vulnerabilities, being able to establish relationships with vendors to disclose vulnerabilities and then implementing remediation.
He defined a vulnerability as “a way for somebody to leverage or use a system in a way that wasn’t originally intended.”
A growing number of software companies are actively looking for vulnerabilities — for example, through bug bounty programs — and more is also being done as a community, Gibson said, but he added that many vendors, particularly smaller firms, remain reticent to work with security experts to fix problems.
Another principal challenge is that managing security over a massive amount of connected devices is difficult and nearly unmanageable, he said.
Meanwhile, some industries, particularly manufacturing and medical, are particularly vulnerable because they often have to
balance downtime of critical equipment
caused by maintenance, including security updates, with productivity and efficiency, he added.
Gibson also highlighted the real-world impact of IoT security, including the rising threats of botnets, destructive malware and ransomware, saying that luckily, for the moment, these challenges are manageable.
Technologies like AI could provide bad actors with a tool to further exploit and profit from vulnerabilities, he added.
“I think we’re going to see some interesting attacks,” Gibson said, but added: “I don’t think, at this point, the bad guys have really figured out how to monetize exploitation of connected devices.”
He issued recommendations for manufacturers, such as leveraging the free energy of the security research community and ensuring that products are updated, while recommending that consumers do their due diligence.