This week, the New York City Council is asking some important questions about artificial intelligence: How do we make sure these technologies are safe? How do we create appropriate guardrails? And how do we understand and report on their impacts?
Those are important questions. But I want to add another one: What does AI safety mean for workers?
Because if we’re serious about protecting New Yorkers from the risks of AI, economic security has to be part of that conversation.
One of the proposals before the Council, Introduction 161, sponsored by Council Member Carmen De La Rosa, would require City agencies to report on how algorithmic tools are affecting the public workforce, including whether positions are being eliminated or workers displaced, how jobs and responsibilities are changing, and where workers need new training.
That matters. But identifying the impact is only the beginning.
For workers, the question isn’t simply whether AI is safe or whether we can measure its effects. It’s more immediate: What happens to me when this technology changes my job?
What happens to the entry-level worker whose foundational, first-rung tasks are increasingly automated? To the mid-career professional watching the skills that built their career shift in real time while being expected to work in fundamentally different ways? Or to the seasoned leader responsible for managing a workforce through changes they’re still learning to navigate?
We shouldn’t wait until workers lose their jobs to respond.
New York City has an opportunity to build worker-transition infrastructure alongside technological guardrails.
That means helping employers redesign jobs responsibly. Giving incumbent workers access to training before they’re displaced. Creating clear pathways into emerging and AI-enabled occupations. And making sure that when displacement does happen, workers have somewhere to turn for the support they need to navigate what comes next.
Our workforce development system has an essential role to play here. But that requires us to think differently about when we intervene.
We cannot build a system that waits for someone to lose their job before deciding they need help. We need to get much better at intervening while the transition is still happening.
We don’t know exactly how AI will reshape New York City’s labor market. And we don’t need to pretend that we do. What we know is that work is already changing. And we have a choice about whether our systems change with it.
If New York City is serious about building guardrails for AI, we should be just as serious about building guardrails for workers.
AI safety can’t just be about making the technology safer. AI safety has to include economic security.