How AI can be a game changer for government

Earlier this year, the union representing U.S. airport security agents called for hiring 6,000 additional screeners to reduce long waits and help handle the rise in summer travelers. But hiring that many people isn't as simple as posting a job ad. Government positions -- especially those at an agency like the Transportation Security Administration -- require a rigorous screening process in which officials might consider 100 employees for every one hire made.

Airport security is clearly an area that could benefit from automation. Through a process known as organizational change management, TSA could automate some entry-level IT jobs that were previously handled by tech employees and repurpose those employees to ease the demand for additional security screeners.

Automation allows employees to focus on more complex tasks, and artificial intelligence-based technology can handle knowledge work in a fraction of the time and with no errors — a level of performance humans can't match.

After decades of research and slow progress, AI is finally growing up. This summer, the White House hosted a series of public workshops on the topic, and the National Science and Technology Council formed a new subcommittee on machine learning and AI to coordinate federal activity.

Although some critics fear that handing over certain tasks to computers would put people out of work, that doesn't tell the full story. Some entry-level IT positions might indeed be eliminated, but employees -- especially those with credentials or security clearances -- could then be repurposed to higher-level work. For many agencies that have a help desk, call center or operations center, customer service would dramatically improve. And any cost savings could be invested into hiring new kinds of workers who can tackle more complicated needs.

Take the IRS: In 2004, the agency answered 87 percent of calls, and taxpayers were on hold less than three minutes on average. But due to major budget and employee cutbacks, by 2015, the IRS could respond to only half of the projected 100 million calls it received between January and April 15, National Taxpayer Advocate Nina Olson found.

It’s clear that a different approach is needed, especially as the Affordable Care Act and other regulations introduce more complexity into our tax system. Smart virtual contact agents can resolve many issues and answer calls more quickly during tax season, leaving IRS employees to focus on the most complicated inquiries.

Likewise, the Education Department runs service desks with only a handful of full-time employees who must handle tens of thousands of requests that come in via phone and email. Virtual contact agents could answer common questions about office hours and eligibility screening, freeing up the human agents to handle complex service inquiries, complaints and questions.

Such technology is already at work in the public sector. In the London borough of Enfield, the local council -- which receives approximately 55,000 phone calls and 35,000 website visits a month -- is deploying a cognitive agent to respond more quickly to inquiries regarding licenses and permits and help residents find information online. The virtual employee communicates with people using natural language, fields the straightforward inquiries and refers tougher requests to human colleagues.

Governments must do more with less, and automation and AI can help solve that puzzle. Let's use technology for what it's always been good at: taking over the routine tasks so our workdays can be more creative and fulfilling. The U.S. government has an opportunity to lead the way by intelligently applying such technology on a bigger stage than anywhere else.