NYC's MyCity Business chatbot, launched October 2023 on Microsoft Azure AI, was found in March 2024 to routinely give illegal advice on housing, tips and cash payments; the city kept it live for over a year despite the findings, before it was shut down as 'functionally unusable'.
Details
Promoter
New York City Office of Technology and Innovation
Period
2023–2025
Keywords
digital government, business regulation, generative AI, public administration
Description
New York City launched the MyCity Business chatbot in October 2023, built on Microsoft Azure AI services, to give small-business owners plain-language answers about municipal rules on hiring, housing and operations. By March 2024, testing by the nonprofit newsroom The Markup — co-reported with THE CITY — found the bot routinely gave confident, incorrect and sometimes illegal advice: it told users a landlord 'is not required' to accept tenants with Section 8 vouchers, when NYC law explicitly bans source-of-income discrimination; it said employers could 'take a cut of your workers' tips'; and it claimed there were 'no regulations' requiring businesses to accept cash, despite a 2020 city law mandating cash acceptance. The OECD's AI Incidents and Hazards Monitor logged the case as a hallucination-driven incident, also noting the bot gave unsafe food-safety guidance in some tests. Despite the reporting, the city's Office of Technology and Innovation kept the tool live, calling it a 'beta' program that had 'already provided thousands of people with timely, accurate answers,' and committed only to future upgrades rather than an immediate fix or suspension. The chatbot remained online for well over a year after the flaws were first reported publicly; incoming mayor Zohran Mamdani later announced plans to shut it down, describing it as 'functionally unusable' and noting it had cost the city roughly $600,000.
Read the full analysis: https://themarkup.org/artificial-intelligence/2024/03/29/nycs-ai-chatbot-tells-businesses-to-break-the-law
Implementation
Implementation detail (cost, timeline, staffing, conditions for success) is not yet available for this practice.
Do you run this practice?
Claim it —
verified implementers get a public contact pathway and can propose corrections.
Data sources
Where this practice's information was retrieved from, and when.
Since April 2019, Canada's Directive on Automated Decision-Making has required every federal automated decision system to complete a public Algorithmic …
The UK's Algorithmic Transparency Recording Standard requires central government bodies to publish structured records of algorithmic tools that significantly influence …