Government robots, chatbots are coming — better define their role now

What do “Star Trek” (original series, of course) and “Transformers” have in common? They have been closer to more accurately predicting the future than all the futurists combined. From Artificial Intelligence (AI) infused computers, to handheld communications devices, to robots that can replicate human tasks, to lethal autonomous weapons systems, fiction has become fact.

One fact I have written about before is how the United States is in a race with China for robotic and AI superiority. The implications cover national security, defense, health care, the economy, education, public safety, transportation, outer space and more. In a couple years it will be easier to name the aspects of our lives that robotics doesn’t affect, instead of trying to list all those parts it does.

{mosads}The slippery slope of robotics and automation is a siren call, fraught with implications of who will be serving who in coming years. One intriguing application of robotics and automation is government services. Instead of worrying about replacement, a more compelling view is to look at how robotics and automation can augment and increase government effectiveness.

According to the IRS and the National Taxpayer Advocate, “In fiscal year 2018, only four out of ten taxpayers calling to reach a live assistor will succeed.” The 2015 Annual Report to Congress had a more depressing observation: “The IRS has steadily decreased the availability of face-to-face assistance at Taxpayer Assistance Centers (TACs), leaving taxpayers with few other options for communicating with the IRS, such as writing a letter or making a phone call.”

The IRS said it is “exploring” chatbots and AI for use with taxpayers. Anytime someone from the federal government says they are exploring, it means not in your lifetime. If the IRS wants to increase voluntary compliance and receive a higher portion of outstanding tax bills, maybe having quality customer service is the first step.

It’s not that most people don’t want to pay what they owe; no one wants the dreaded IRS letter. It’s that they don’t want to celebrate their next birthday on hold waiting for someone to help.

Google’s demonstration of Duplex, “An AI System for Accomplishing Real-World Tasks Over the Phone,” could be the customer-service killer app the government has been waiting for. The all-too-real voice fooled everyone on the other end of the line. What if I could call the IRS at 3:42 AM and get a question answered? I’m not sure I would care if it was AI. What I care about is that I got an answer on my terms, not theirs.

The Dubai Police in the United Arab Emirates are deploying robots equipped with AI to address the basic concerns of citizens and visitors. These life-sized constables are multilingual and can allow you to pay fines or report crime.

In fact, these mechanized officers will allow you to poke them in the chest without going to jail. Instead of handcuffs, a touchscreen with additional resources pops up. Police officials expect 25 percent of the workforce to be robotic by 2030, but not at the cost of real officers. This is a concrete example of an augmentation strategy that benefits everyone.

Another area screaming for improvement is airport security. Not a year goes by without a report from the inspectors general talking about how Transportation Security Administration screeners failed to catch explosives or guns. They did, however, prevent a 96-year-old woman in a wheelchair from perpetrating any dastardly deeds. (Full disclosure: I have had more than one water bottle confiscated.)

According to a Washington Examiner article, in June of 2015 “The internal tests were conducted by Homeland Security Red Teams, undercover investigators who pose as passengers with harmful objects hoping to get through TSA security checkpoints. According to a recently released Homeland Security Inspector General report, Red Team members were able to get through security checkpoints with potential weapons in 67 out of 70 tests.” That’s a 95 percent failure rate.

That wasn’t the only time. Right before the July 4 holiday on 2017, the same Homeland Security Red Team got 16 out of 17 potential weapons past the TSA screeners in Minneapolis. That’s also a 95 percent failure rate.

You can’t hire your way out of a mess like this. The time it takes to recruit, review, screen, conduct background checks, and then hire and train, is daunting. Robotic screening and AI could accelerate the process, leaving highly-seasoned screeners to review anomalies and actually catch guns and explosives faster, with better outcomes.

Another major area of improvement is privacy. No one wants the government collecting more information than absolutely necessary. The protection of PII — Personally Identifiable Information — is more important than ever in our always-on, always-connected society.

Tasks that would normally expose PII to humans can now be managed by AI. While a data breach is always a consideration, a significant amount of the breaches today are the result of a malicious insider.

The future is unpredictable, constantly evolving, and changes on a whim. What is predictable is that robots and AI will be a significant part of it. The sooner we embrace it, and define the proper role for this technology, the faster we’ll get to the point of where AI can tell you the 96 year-old grandma isn’t a real threat.

Morgan Wright is an expert on cybersecurity strategy, cyberterrorism, identity theft and privacy. Previously Morgan was a senior advisor in the U.S. State Department Antiterrorism Assistance Program and senior law enforcement advisor for the 2012 Republican National Convention. Follow him on Twitter @morganwright_us.