Congressional Briefing on Artificial Intelligence for Policy Implementation
Mar 9, 2020
Carnegie Mellon University held a briefing this week to discuss artificial intelligence (AI) design for policymakers and policy implementation. The panelist discussed the use of AI throughout the world and way in which it is currently being used. Crystal Cody, Public Safety Technology Director, City of Charlotte, NC, shared how she and her team are using AI to track their police and public safety officers. Through their systems, the police officers are monitored for safety and are flagged when there a safety concern such as the use of force or use of a firearm. The system sends the alert to the head officer for review so the appropriate action can be taken. Cody and other panelists stressed how efficient this system has been for their force and recognizing officer that may need their help.
Overall, all the panelist stressed how important it is to not eliminate the human factor when using AI. Rayid Ghani, Distinguished Career Professor, Machine Learning Department and Heinz College of Information Systems and Public Policy, Carnegie Mellon, stated, “machines are terrible at commonsense reasoning”. Andrew Fano, Artificial Intelligence Managing Director at Accenture Labs, added that machines should be used to spot things that require human attention. The machines can pick up on things humans may miss and send an alert for review. Ghani stressed that one of the biggest issues currently facing the implementation of AI is that it is too focused on building autonomous systems that are not designed to communicate with human. He mentioned how this is skipping a step and is leading to many issues of distrust and biases.
The panelists also highlighted some of the biggest barriers and obstacles currently facing the implementation of AI from both a policy side and local side. Some of these barriers and obstacles include:
- Local governments do not know what is possible with the use of AI and do not have access to people who are experts in that field.
- Even if local governments can implement AI, they must start from scratch.
- There is a lack of trust and policies for implementing AI.
- Many local governments do not have a budget that would allow them to hire people who have expertise in AI systems.
- There are also not enough people who have the skills and trainings to operate AI systems.
- The overall lack of people across disciplines working together to create better systems.