One year ago, New York City introduced a sweeping AI plan. Here’s how it’s going.

As states and cities wrestle for a leadership role in artificial intelligence, New York City made one of the biggest splashes last year as it unveiled its sweeping AI Action Plan.

The ambitious roadmap laid out 37 “key actions” to be taken across a range of areas, 29 of which had to be started or completed within 12 months. Those actions include establishing an AI governance framework, creating an external advisory body, and engaging with the public and city employees.

The rollout hasn’t been without its issues. Most notably, the city garnered negative headlines earlier this year following the launch of its much-vaunted MyCity chatbot. The bot was supposed to give New Yorkers information on starting and operating a business. Instead, it told bosses that they could take workers’ tips and landlords that they could discriminate based on source of income. Those issues have since been resolved, according to officials.

“No surprise,” New York City Chief Technology Officer Matthew Fraser said this week during the Google Public Sector Summit in Washington, D.C. “When you do anything that's new, you will get some feedback.”

But despite some hiccups, city officials say they are on track or ahead of schedule to meet the goals laid out in their action plan. In fact, they said, they are already looking ahead to how AI can help in areas like procurement, and how to get every agency fully up to speed on the role the technology can play to help their work.

Route Fifty sat down recently with Alex Foard, executive director for research and collaboration at the New York City Office of Technology and Innovation and one of the action plan’s lead architects, to look back and to talk about what comes next for the city and its AI efforts.

This interview has been edited for clarity and brevity.

Route Fifty: How have the last 12 months been since the unveiling of the AI Action Plan?

Alex Foard: It's been great. We really wanted to focus on a couple of different things. We wanted to be clear that the city can and should be using AI. It’s something that we've been doing for a long time, something that we think can really add value and improve the way that agencies are doing business and serving New Yorkers. The second component is that we do think that we need to be using AI safely and responsibly, and so we want to be able to commit to building out the framework and the structures that we need to make sure that responsible use is the approach.

We also wanted to focus on the opportunity to be constructive and intentional with our steps, and focus on real, concrete actions instead of just abstract thinking. That's how we landed on what we did. When you look at the 37 actions that we have in the action plan over seven different initiatives, it really reflects a wide range of commitments for the city. We are focusing on building out those or taking on those initiatives that we've laid out for ourselves. Of the 37 key actions, we had committed to at least initiate, if not complete, 29 of them within the year. We are on track to slightly beat that target, so we're pleased with that progress.

We've been able to highlight a couple of different areas and really make some progress on governance, building out some of our initial policy, building out our internal AI steering committee, which is helping to advise on policymaking and other decisions for the city's AI. We’ve been able to hold some public engagement sessions, chatting with New Yorkers about their AI priorities, and hearing from them what they think about some of the work that we've been doing. We've been able to build out a lot of the technical resources that we need to be able to better support agencies in their work.

We've been thinking a lot about skill building for city workers, how we are equipping city employees to be able to work with AI tools and understand the role that AI tools play, but also thinking a lot around the procurement support or the in-house development support that agencies may need to help them think through bringing AI tools online.

It's been a busy year. We have a lot of work still ahead of us, but we're really happy with the progress that we've made.

Route Fifty: Has it been challenging? It's a lot of boxes to check. How do you go about tackling something like that when you give yourself a big to-do list?

Foard: We're looking to create a meaningful sequence of steps that the city can take to help support this work. In many ways, we build upon each step as we go. As we do more analysis to help us understand, for example, where skills gaps may be needed in the workforce, we can then say to ourselves, “What resources do we currently have available? What additional resources do we then need?” It is a lot of plates spinning, certainly, at a given time, but there's very much a sense of a unified theme across all of the work that we're doing, and that helps to keep us targeted and oriented around the outputs for individual actions and then the overall objectives for the plan.

Route Fifty: What do you know now about AI that you didn't know 12 months ago?

Foard: Well, it may not be so much what we didn't know. I think we definitely have been, as everybody has, really marveling at the role that AI has been playing in the national conversation, becoming a term that people are using on a day-to-day basis, not just in technical circles. The evolution of technology is really bringing this to the top of mind for everyday New Yorkers. That's been something that we've wanted to be responsive to, making sure that New Yorkers are equipped to not only understand what AI may mean, but also what sort of impact that has on the way that their government does its work.

Route Fifty: You mentioned the governance and the policy side of this. What's going into that?

Foard: We started at the bottom with the most important pieces. We published our principles and definitions guidance earlier this year, and that lays out the city's major principles for how we want to be using AI, making sure that we're thinking about core components like privacy and cybersecurity, but then also thinking about appropriateness and the sustainability and fairness of the AI tools that the city uses.

Then, really importantly, definition setting is critical to help not only city agencies, but engaging with the public to better ensure that we're all speaking the same language and talking about the same things. We then also prioritized some preliminary use guidance around generative AI tools for city agencies, recognizing the impact that branch of technology is having on the public conversation. As we continue to think through governance over the next year, we really want to build out more of the policy commitments that we have in the action plan, really focusing on risk assessment, but then also focusing on areas where we want to support agency contracting, agency procurement, and where procurement becomes a tool for helping us to reiterate our values and principles as a city.

Route Fifty: The other thing that struck me was about the public engagement, you guys going out into the community and talking to people about this. How would you characterize how New Yorkers are feeling about this? Are they nervous? Are they excited? Somewhere in the middle?

Foard: Not surprisingly, it's a complex conversation, and a lot of people have differing priorities when it comes to thinking about AI. Our goal with public education and public engagement is about meeting people where they are from a language and technical point of view, making sure we can all speak the same language and understand what we're talking about when we talk about AI issues, but then also hearing specific groups’ concerns and priorities. We do expect that to be different from one forum to the next. We're open ears when it comes to hearing what is top of mind for people. But in general, one of the encouraging things is that people are talking about it, and we're happy that people are having conversations and asking a lot of interesting questions about it.

Route Fifty: Do you get the sense that employees are experimenting with AI right now? How are employees approaching it right now?

Foard: The city's been using AI for a long time before the advent of generative AI, and we see that really reflected in the public reporting that we do. The city does yearly public reporting around its algorithmic tools. A lot of those are AI tools, and you can really see in that report just how important a lot of these AI tools are for agency business. You really see these tools integrated into the way that agencies are supporting their decision-making.

We know very well that certain parts of agencies are well-versed in these technologies and these techniques. The interesting thing over the next year or so, part of our commitment through the action plan, will be meeting other parts of the business, to be able to speak to other components of the city's workforce, to make sure that they're acquainted with AI tools, that they know how they work, what sort of impact they can have in the other parameters that lead to safe and responsible use.

Route Fifty: I wanted to ask about training, upskilling and that sort of thing. What do you think that is going to look like for your employees?

Foard: It all starts with level setting around terminology and familiarity with concepts. You hear a lot of different things described as AI, particularly these days, so it's important for us to make sure that we are all talking about the same set of things. We use an expansive definition of AI here for our policy work, to make sure that we're capturing a wide array of technologies that can fit into that bucket. We really, first and foremost, want to make sure that we have that shared language, and that we know that people are understanding what is meant by AI and not focusing just on what may be the most readily available piece of information about it. The key thing to think about is the opportunities to work with AI, but also to work on matters related to AI, and that's something that we're particularly mindful of: Making sure that city employees can think about it from multiple perspectives.

Route Fifty: When the MyCity chatbot for businesses came out it had some difficult headlines in its early days. What have you learned from that whole experience?

Foard: I'm going to have to defer some of that to other parts of the New York City Office of Technology. And just to be clear, my team is not a business owner of the MyCity chatbot, so I'm not able to speak to some of the particulars of the system. I will say, I think what's important is two things: One, that we are encouraging innovation for city agencies, focusing on opportunities to bring innovation to the forefront, and being able to try new things. Two, that our position is clear that governance and management of AI tools is important. The governance piece is there to make sure that as we innovate as a city, we're doing so in a responsible way.

Route Fifty: How do you measure the success of AI’s rollout in the city?

Foard: That's a really interesting question, and I don't think there's a singular answer because agencies use AI tools in really different ways that reflect the particularities of their business. So, we always want to be responsive to the individual agencies' needs and how they're meeting their own objectives and goals. It's great to see them use these tools in successful ways that really are able to create new efficiencies for them, or to optimize some of the resource allocation that they've been using in some cases to keep New Yorkers safe in differing ways.

At a broader level, we're thinking about our opportunity to follow through on our action plan commitments, thinking through where we're building out the policy and the guardrails that we've committed to, but where we're also building out the processes to support agencies as they're doing their work so that they can then meet those New Yorkers’ needs more directly.

Route Fifty: What's next?

Foard: Happy as we are with the progress we've made on the action plan, we still have a few things left to do. There’s still governance that we are planning to build out, in particular focusing on risk assessment over the next year. We’re interested in supporting contracting and procurement, and how we can also use procurement as a tool to reflect our own values. Staff training is a key component here, making sure that we're bringing those resources into the city's workforce.

Then outside of the action plan, as we move forward, public engagement is going to continue to be a major focus for us to expand on the engagement that we've done so far, with additional opportunities to connect directly with New Yorkers. Then also really keeping pace with the change in technology. We must make sure that as we are doing this work and as we are completing the actions from the action plan that we're reflecting on how the technology is changing and what we need to do to keep our policy and our work current.