Mayor Eric Adams has unveiled a plan for New York City government to roll out usage of artificial intelligence. Although there is reason for some skepticism, the plan offers important concepts that every state and local government should consider.
AI and local government
Most attention on the economic impacts of AI has focused on business, but government will also have widespread opportunities for productivity gains. The large language models such as ChatGPT, Bard and Claude, will raise the productivity of workers across a wide range of tasks. That will enable an organization—including government—to do more work with fewer employees. A Goldman Sachs analysis puts at the top of the AI-impacted list: office and administrative support, with high scores also for legal; architecture and engineering; business and financial operation; and community and social services. This covers a wide range of government office jobs. At the other end of the spectrum, with low AI impact, are maintenance, construction and manufacturing production. Police and fire fighting are not identified separately, but they would no doubt be very lightly affected by AI.
AI for city permits and regulation
Mayor Adams put high emphasis on helping companies cope with city permits and regulations: “… MyCity Business will connect business owners and aspiring entrepreneurs to content to help them start, operate, and grow businesses in New York City. With the AI chatbot, business owners will more easily be able to access trusted information from more than 2,000 NYC Business web pages.” This is a great use of artificial intelligence. Though some of us would like to see permitting and regulations simplified and cut back, the AI approach is more likely to be enacted rather than just promised.
AI risks in government
New York’s plan also emphasizes dealing with risks. “The New York City Artificial Intelligence Action Plan introduces a set of phased actions the city will undertake to help agencies evaluate risks and determine whether or not a tool is the right technology to deliver better positive outcomes for New Yorkers.” As noted in my article about which companies would use AI, about half of the early business policies simply forbid ChatGPT. Business concerns emphasized false statements (called hallucinations) from LLMs, and exposure of confidential information used in prompts. New York’s policy announcement adds the potential for bias and discrimination. These are reasonable concerns.
Politics of AI in government
A government choice to use AI will be a political decision. Predictably, the city will “Create an external advisory network to consult with stakeholders across sectors around the opportunities and challenges posed by AI.” This no doubt includes public employee unions that will be concerned about the number of jobs and potential layoffs. Other interest groups may also have worries about AI, though some should applaud the potential for better customer service.
AI is still in its early stages. Most of the usage will not entail people asking ChatGPT to answer questions or write memos. Instead, we are seeing many start-ups developing specialized applications, to solve specific problems, using a large language model in the background. The user won’t notice the AI directly, but will simply get results that help to perform a task. Some of these apps will have use in government as well as the private sector. Think about performance reviews, bookkeeping and purchasing, which are functions common to both the private sector and the public sector. As more governments become comfortable with AI-powered tools, other companies will develop specialized applications specific to government operations.
AI for better city services
Just as consumers typically want better quality at lower price, citizens generally want better civic services with a lower tax burden. Competition leads businesses to employ tools that can help them deliver more for less. Without competition among governments, citizens must step up. They can begin by holding up to their cities and states the New York plan, asking if their local government is also working on a plan.
On the spending side, department managers at budget hearings could be asked how they will use AI to reduce staffing needs. The aggressive form of the question would ask how many positions could be cut; the milder version would assume that productivity improvements would eliminate the need to replace retiree for the next 10 or 20 years.
Jokes about poor customer service often include the department of motor vehicles. Citizens can ask how AI will be used to respect the public’s time when they are complying with all regulations and licensing requirements.
Safety is also a legitimate concern. Citizens could ask if humans can be contacted in case of AI-generated errors, how their private information will protected and whether bias has been minimized in AI solutions.
In short, people who pay for and use government services will ultimately drive decisions about public use of artificial intelligence.
Read the full article here
