PUBLIC TESTIMONY
Testimony by Thomas Gilbert, Founder and CEO of Hortus AI, on AI and Surveillance Technology Oversight
4:08:19
ยท
3 min
Thomas Gilbert, founder and CEO of Hortus AI, testified on the importance of public oversight for AI and surveillance technologies in New York City. He drew parallels between historical urban planning approaches and current AI implementation, emphasizing the need for transparency, regular audits, and community involvement in technology oversight.
- Gilbert supported three proposed pieces of legislation (Intro 168, Intro 233, and Intro 480) that would enhance NYPD's accountability in using surveillance and AI technologies.
- He highlighted Hortus AI's mission to empower local communities in assessing and integrating AI technologies.
- Gilbert advocated for more intensive oversight of emerging AI technologies, such as chatbots, and offered to work with city officials on progressive audit frameworks for AI systems.
Thomas Gilbert
4:08:19
Good afternoon, members of the New York City Council.
4:08:22
My name is Thomas Gilbert.
4:08:24
I am the founder and CEO of Hortus AI.
4:08:27
I'm here representing Hortus' mission to empower local communities to assess and integrate AI technologies on their own terms.
4:08:36
So precisely one hundred years ago, Robert Moses set up shop at 302 Broadway overlooking City Hall.
4:08:42
And through a combination of graft incentives, cutting and deceit, most of it legal, Moses rebuilt New York City in his own image under the aegis of public safety, and he did it largely through surveillance.
4:08:55
Opposing Moses, Jane Jacobs wrote that a street needs three things in order to be safe.
4:09:00
First, a street must have a clear separation between public and private.
4:09:05
Second, it must have the watchful eyes of storekeepers, residents, and those passing by.
4:09:10
Third, a sidewalk is needed so that people can use the street regularly even without cars.
4:09:16
In other words, what that history tells us is that streets are not made safe by technology.
4:09:21
They're made safe by having certain clearly defined properties, namely as public, as watched, and as regularly used.
4:09:29
Moses thought surveillance could make the public safe.
4:09:31
Jacobs knew it was the other way around.
4:09:34
Today, the Committee on Technology is considering how to more clearly implement the public oversight of Surveillance Technology Act.
4:09:40
The three the three pieces of legislation introduced by council members Farias, Crystal Hudson, and Julie Wan are not just good ideas.
4:09:47
They respectively enact Jacobs' three criteria for safety.
4:09:51
One six eight would require the NYPD to evaluate and report on the private surveillance technologies it uses for public benefit.
4:09:58
Two three three would require regular written audits of the NYPD's use of facial recognition technology and to widely share the audits findings.
4:10:06
And four eighty would ensure continuous transparency in NYPD's required impact and use criteria.
4:10:12
These proposals reflect a growing awareness that AI technologies are not safe because they can learn from data or recognize faces or managed by technocrats.
4:10:21
Rather, AI technologies are safe because of their commitments to and from public interests.
4:10:27
Emerging forms of AI such as chatbots will require even more intensive forms of oversight, regular audits, and substantive transparency.
4:10:34
Fortis' work is designed to address this.
4:10:37
Alongside our work with the public sector, Fortis has outlined what is toxic about AI today, namely prioritizing business objectives over quality of life and its impacts on communities and how AI could instead be built differently.
4:10:49
Cortis solves this by providing tools to government entities to build AI and implement it for active citizens in partnership with local institutions.
4:10:59
In tandem with proposed office of algorithmic data integrity, we hope to work with New York City Officials and pose more progressive audit frameworks for AI systems from facial recognition to generative AI applications.
4:11:10
I invite members of the city council and my fellow citizens to join in this work to ensure that oversight of AI systems is of, by, and for the people.
4:11:20
Thank you for your attention.