NORTH ADAMS, Mass. — The City Council is considering ways to control the use of so-called artificial intelligence for public services. 

 

The draft ordinance is proposed by Council President Ashley Shade, who said she has been working for more than a year on language for a new chapter in the city's code — "Safeguards for Artificial Intelligence, Fairness & Equity."

 

"The language that I proposed was from a mixture of different ordinances that other communities have adopted, but there are no ordinances like this in the commonwealth, or even in this country, that I could find," she said at Tuesday's General Government Committee meeting. "I built this so it could be built upon. The whole point of the language in here is so that it's a starting point, and that it gets continually built up. ... 

 

"The number one thing that this ordinance does, and the most important thing to me about this ordinance, is that we are protecting the civil rights of the people in our community."

 

Shade, a member of the committee, told the dozen attendees at the meeting that AI was happening; but the city could regulate it and require it be used in a responsible way. 

 

The AI Safety ordinance basically defines two types of AI: high risk and low risk. Low-risk are applications and software that hold no decision-making capabilities such as for transcription, spell checking, etc. So internal administrative, clerical, or productivity tools that "do not materially affect rights, benefits, or enforcement outcomes shall not be considered high-risk."

 

High-risk is any application being used for public services that could 1) affect someone's legal rights, benefits or access to services; 2) employment decisions such as hiring, evaluation, discipline or termination; 3) code and law enforcement; 4) surveillance, monitoring and tracking; and 5) that present a risk of discrimination or "disparate impact under applicable law."