Verenigde Staten van Amerika - The New York City AI Strategy

Voor de volledige samenvatting, zie het ENG-gedeelte

What: Policy orienting document

Impact score: 3

For who: Government, business, non-profit organisations, citizens

URL: The NYC AI Strategy

National AI strategies often set out a structure with a concrete series of actions. This strategy is rather the result of a thinking exercise and provides a set of proposals for action. It is up to the city council to decide whether the proposals will be implemented in concrete terms.

The AI strategy is part of a series of plans that NYC has rolled out. It has already introduced an IoT strategy and an Internet Master Plan. New York's strategy is quite broad and mainly seeks to build on existing structures. It also wants to make maximum use of the innovation ecosystem present in the city and the opportunities it offers.

The NYC AI Strategy has three main components:

  • It establishes a baseline of information about AI to help ensure decision-makers are working from an accurate, balanced, and shared understanding of the technology and the issues it presents.
  • It outlines key components and characteristics of the local AI ecosystem today, highlighting the wide range of institutions and efforts in place in New York City as well as providing insights into City government and how different offices and agencies are using or play a role in the governance of AI.
  • It identifies five broad thematic areas in which the city should focus further as part of an approach to AI:

    • City data infrastructure
    • AI applications within the city
    • City governance and policy around AI
    • Partnerships with external organizations
    • Business, education, and the workforce

City data infrastructure

Current state: NYC’s existing approach to data is wide-ranging, if fragmented, and can be built upon to support both AI applications and other important improvements. Within city government, this approach includes both a city-wide component as well as plans at the level of individual agencies or key domain areas like health, education, and transportation.

There have additionally been a variety of efforts to bring in data from external sources to support city goals. For example, private sector companies in the taxi and limousine business are required to share the data on their rides. The purpose of this measure is to inform the city’s transportation policymaking.

The use of data sharing agreements is an important part of the city’s overall data management infrastructure. One key example is the leadership of the Chief Privacy Officer and the Mayor’s Office of Operations in creating the Citywide Data Integration Agreement (CDIA), which is a master framework agreement that all city agencies have signed, which specifies many of the required privacy, data security, and other terms appropriate for multi-agency data sharing agreements.

Opportunities identified by the strategy: Firstly, the fragmented approach must be addressed. Therefore, it is suggested that a citywide data strategy should be created, and that consistency be built into agency-specific strategies. In addition, common agency pain points in relation to data should be listed. There are a number of issues around data that are common across agencies and would benefit from a more centralized approach, rather than requiring agencies to build this expertise in-house. A centralised guidance to assist agencies should be created for this purpose.

A centralised approach is the first step, a deepening of skills is the second. It is suggested, for example, that more data engineering (the work to make data usable) expertise should be cultivated across government agencies.

Lastly, the strategy proposes to modernize the city’s data infrastructure. There is also work to be done in the field of open data: the standards must be improved to make it more reliable and usable. Better documentation should help to ensure that datasets are presented in context, are less likely to be misinterpreted by users, and can help facilitate ethical use. Continued efforts to improve compliance and documentation standards, as well as creating more links between agency datasets, will save a significant amount of time and effort and help users unlock more value from the city’s public data assets. Improved guidance by the city (indicating what can be relevant for which topics, which data fit together, etc.) should boost the use of open data for private tasks and problems.

AI applications within the city

In this chapter, a number of opportunities for making good use of AI in the city are reviewed. It also looks at how the city administration and its agencies should approach the process of implementation and what the needs are.

City governance and policy aroud AI

The city government wants to provide robust transparency on the internal systems it will use (explaining how it works, how accurate it is, how performance is evaluated). Furthermore, a structure will be worked out to make ongoing, regular review possible.

The government will also set up models for community engagement and participatory approaches to innovate in public participation in the design, use, governance and policymaking related to AI systems. It can build therefore on existing programs and partnerships. This can include broad-based public education efforts, fostering participatory approaches to use case identification, engaging the public throughout system engineering, or other activities.

Furthermore, the city government should develop a policy for procurement of datasets and AI technology. This involves building and maintaining institutional expertise and offering tools to support rigorous evaluation of vendor claims about their use of AI. The position of Algorithms Management and Policy Officer is already created in the administration. It is proposed to extend the range of duties. This staff member would then perform tasks such as creating a policy framework and set of management practices centred around fair and responsible use of algorithmic tools by city agencies.

Lastly, it is proposed that the city government should promote experimental and empirical policymaking. It should establish an internal working group to support a holistic and coordinated approach to AI policymaking and procurement, consisting of both city agencies procuring AI and those focused on technology policy. This working group can iterate policies and protocols, elevate best practices, and seek input on emerging matters of shared concern.

Partnerships with external organizations

Within the walls of city departments, New York wants to build internal staff capacity to assess, use and work with AI. Additionally, the government will expand its capacity to use and plan for AI by fostering partnership opportunities with external experts from its rich local ecosystem. Forging and managing these partnerships should be eased by central coordination.

It is suggested that permanent structures of engagement be established between various stakeholders in the city's AI ecosystem and agencies of the city government. The city can act as a convener within the local ecosystem — gathering researchers, civic tech and advocacy groups, industry practitioners, and community groups to share work in a bi-directional way, and discuss emerging issues and opportunities. The models the city will use should emphasize interdisciplinarity, equity, and inclusion. It is also recommended that a team be appointed to link the needs of the city's agencies with academic institutions that can offer support.

Business, education and the workforce

The city government will further develop a vision on how it protects the digital rights of consumers, workers and small businesses. It also plans to set up projects on the impact of AI on local workforce and other AI-related concerns about the future of work. The city government also wants to anticipate this development. It will design programs that are responsive to evolving industry needs and practices, anticipate and plan for job and skill displacement, and effectively address equity and underrepresentation in the local AI workforce.

New York has set up a structure, The City’s Tech Talent Pipeline (TTP), which engages with local tech industry stakeholders to understand their workforce needs and establish an inclusive pipeline of New Yorkers equipped to fill them. This includes developing training programs, working to increase the number of university’s Computer Science graduates, and connecting New Yorkers to tech internships and jobs.