IoT Predictions for 2019: Less Hype, More Pragmatism

Efforts to Regulate Facial Recognition Technology Gain Ground

In recent years, facial recognition technology has won mainstream status. Modern phones such as newer iPhones or Samsung Galaxy phones offer the option to unlock the device with a user’s face. Airports such as Washington Dulles International are using the technology to confirm the identity of passengers.

The technology is also quickly expanding for surveillance applications. In China, the technology is de rigueur in many areas. With more than 200 million in use, China uses facial recognition technology to help cement its authority. For instance, it uses the technology to track the Uighur Muslim minority. The Economist notes that other nations with an authoritarian bent could use the technology for similar means, while also acknowledging the potential for First World nations to abuse the technology.

On a related note, the United States Secret Service is investigating the use of facial recognition at the White House to scan the faces of people in the vicinity to determine if they are a “person of interest.” As the ACLU recently wrote, the test program “crosses an important line by opening the door to the mass, suspicionless scrutiny of Americans on public sidewalks. That makes it worth pausing to ask how the agency’s use of face recognition is likely to expand — and the constitutional concerns that it raises.”

While the privacy-related concerns surrounding the technology are clear, so are its potential benefits for defined applications. Gartner expects that, by 2023, AI-enabled facial recognition technology will lead to an 80 percent drop in missing people in advanced nations compared with 2018.

The swift expansion of the technology is drawing warnings from groups such as AI Now, a New York University–affiliated group with membership ties to Microsoft and Google. AI Now concludes that “Governments need to regulate AI by expanding the powers of sector-specific agencies to oversee, audit and monitor these technologies by domain.”

Expect such sentiments to pick up ground in 2019, as the privacy-eroding ramifications of the technology become more manifest. The U.S. states of Illinois and Texas already have restrictions in place regarding facial recognition.

Another factor that could also drive regulation is the fact that some facial recognition systems have also struggled with inaccuracy, with some systems demonstrating high error rates when analyzing individuals with dark skin and women, for instance.

As facial recognition technology advances, it will likely become increasingly used “to detect things such as personality, inner feelings, mental health and ‘worker engagement’ based on images or video of faces,” reads part of the AI Now report. “These claims are not backed by robust scientific evidence, and are being applied in unethical and irresponsible ways that often recall the pseudosciences of phrenology and physiognomy.”

The increasing attention facial recognition technology is receiving is likely to spur efforts in democratic nations to regulate it. Whether such efforts become law in the following year is considerably less certain.

Appeal of Prescriptive Maintenance Grows

Imagine you had to drive from San Jose to San Francisco in the evening on a weekday — any weekday. The 50-mile stretch between the two cities is a 50-minute drive in the middle of the night. But if you left at 4:40 p.m., the drive could take anywhere from one hour and 40minutes to two hours and 40 minutes, based on Google Maps data. But just knowing that the trip will likely take two hours or more may not be the most-helpful piece of information. What if your are running low on gas — where is the best place to fill up without getting stuck in even more traffic? Or if you are hungry? Where might be a good spot to escape the traffic and how long should you plan to spend there to wait it out. Imagine an app that takes into consideration your precise condition, and prescribes options you can take based on your particular needs.

This example provides a rough idea of the benefits of prescriptive rather than predictive analytics. In the industrial realm, predictive maintenance helps organizations know when an asset might fail. Prescriptive maintenance goes a step further by providing advice on what to do to achieve a selected outcome. As Frost & Sullivan explains: “Unlike predictive maintenance, prescriptive maintenance is not limited to merely predicting the failure – it is a strategic maintenance process that allows for the application of the solution, as and when it is needed.”

Sastry Malladi, chief technology officer of FogHorn acknowledged that adoption of predictive maintenance is still new for many organizations, early adopters will embrace the concept of prescriptive maintenance next year, he said. “For example, elevator manufacturers want to put an end to routine problems, such as friction in elevator doors,” he said. A predictive maintenance approach to addressing this challenge is to use edge computing to analyze sensor data at its origin. An organization using the approach can “schedule service before anomalies impact performance in a highly efficient manner,” he said. “As prescriptive maintenance becomes available, before the manufacturers roll a truck to provide maintenance on an elevator, they will have data available to suggest areas most likely to need repairs and have verified the repair staff person the expertise, tools and parts available for the repair.”