
Durkan calls for a hearing on AI programs, in speech written by AI program.
As she asked her colleagues yesterday to approve a hearing on the role of artificial intelligence in municipal government, Councilor Sharon Durkan (Back Bay, Beacon Hill, Fenway, Mission Hill) occasionally stumbled over her words.
She apologized, saying she hadn't actually written the remarks she was reading calling for a hearing - she'd asked an AI program to write them for her.
"Artificial intelligence is no longer some futuristic idea," her computer-driven statement began. "It's here shaping how we work, communicate and govern. From health care diagnostics to city planning, AI has the potential to make public services faster, smarter and more accessible. But like any powerful tool, AI has to be thought of thoughtfully."
Durkan said other cities have already begun to use AI for everything from improving traffic-light cycles and ambulance responses to monitoring Twitter for posts by sick people to stamp out food-borne disease outbreaks earlier. But she acknowledged that any AI discussions will have to also take into account privacy and racial bias - both Boston and San Francisco prohibit the police use of facial-recognition technology - and the fears that the city will just install banks of AI processors to eliminate workers.
"Not all AI is good, not all AI is helpful," she said, pointing in particular to "generative AI" of the sort where you ask questions to "a thing that's having a delusion," she said. "Generative AI is differant than all the other things we can unlock with [AI]: Traffic signals, public health."
At its best, AI is a tool to help people make better decisions - such as members of municipal unions, who she said need to be part of the discussion. And that discussion needs to be "making sure AI is just a tool and not a substitute for human judgment." She said she didn't just ask whatever AI tool she used - she didn't specify - to write her speech; She first had to organize possible topics for it to research and then review and refine what it spit out.
Durkan added that the city should also call on local experts for help - she said researchers at Northeastern and Boston University in particular have been at the forefront of looking at the ethical and privacy issues related to AI.
City Council President Ruthzee Louijeune assigned Durkan's hearing request to the council's committee on city services and information technology to schedule at least one hearing on the issue.
Watch Durkan make her hearing request:
Like the job UHub is doing? Consider a contribution. Thanks!
Ad:
Comments
Which "AI"?
By Kaz
Thu, 03/20/2025 - 12:01pm
AI isn't some "futuristic idea" because we've had machine learning algorithms since some of the earliest punchcard computers. The only reason "AI" seems new is because of the latest Large Language Models that have hoovered up all the text on the internet (regardless of copyright) and are able to then spit it back to us in response to prompts it can interpret (ask something outside its knowledge base and it's clueless at best or will lie to you at worst). They made an impact because they can "talk" to us...but what they say is of limited intelligence and utility.
For example, AI wrote her speech but she didn't give it enough details and/or it has no knowledge of the news from the last 7 months...because Project Green Light (Google Research's AI approach to improving traffic light cycles) started in 2020 (not new) AND just 7 months ago announced with the City of Boston that it's going to work on our data with the city (but she mentioned how AI can help with traffic lights...but not that the city is already doing that with Google right now)!
So, like, the cat is outta the bag at city hall....Google is using AI to help with our traffic lights. So, *now* we need a hearing? Uh...keep up please.
AI is already hard at work in all sorts of ways...it's just that companies have branded these generative AI chatbots as "AI" and now the public thinks it's some new and cool thing that has to be studied...even though decades of work got us here and is already in place in numerous ways.
The term "artificial
By xyz
Thu, 03/20/2025 - 4:40pm
The term "artificial intelligence" has been a marketing term from its invention: the professor at, if I remember correctly, Dartmouth, thought that ARPA would think it sounded cool and fund his grants. As you probably know, but others may not, in modern parlance "AI" encompasses a huge variety of software and techniques, some of which work pretty well, some of which, eh, not so well. More recently people use it specifically to refer to LLMs, "large language models", also known as "stochastic parrots" and "spicy autocomplete".
"spicy autocomplete"
By Don't Panic
Thu, 03/20/2025 - 11:20pm
That's wild!
Is our city council …
By Lee
Thu, 03/20/2025 - 1:58pm
… being run by AI now?
Or the same old Natural Unintelligence á la Ed Flynn?
Got a big hammer
By Plen-T-Pak
Thu, 03/20/2025 - 2:01pm
And looking for nails.
More like ...
By SwirlyGrrl
Thu, 03/20/2025 - 8:34pm
I have a hammer that is slightly improved over the last 40 years, but I have an even bigger hype and marketing department for my UBERHAMMER!
Durkin Should use more AI
By Username Unknown
Thu, 03/20/2025 - 9:31pm
When Ed Flynn seems more literate than Durkin you know you have a problem. Durkin struggles to put two sentences together without an um or you know included and then somehow turns every event or story to be about her.
Add comment