New York Subway Ads Warn of Slow AI Skill Loss

A nonprofit campaign in NYC subways argues that over-reliance on AI erodes human skills, urging users to reconsider personal tasks.
Key points
- The Center for Humane Technology launched a subway ad campaign in New York City to raise awareness about AI drift.
- AI drift is defined as the gradual erosion of human skills and relationships as AI takes over daily tasks.
- Experts warn that the frictionless design of AI tools encourages users to outsource personal and creative work without realizing the long-term cost.
New York City subway riders are encountering a new type of advertisement that moves away from product hype. Instead of promoting new features, these ads carry a warning about a gradual loss of human capability, urging commuters to pause and reflect on their daily dependence on artificial intelligence tools.
The campaign, launched by the Center for Humane Technology, introduces the concept of AI drift. It argues that as AI takes over more routine and personal tasks, people risk losing essential skills, meaningful relationships, and a sense of purpose, creating a subtle but significant risk to human well-being.
Defining the gradual skill erosion
According to Camille Carlton, senior director at the Center for Humane Technology, AI drift is the slow degradation of human abilities as machines handle more of daily life. This is distinct from technical model drift, which refers to a machine's accuracy declining over time. Here, the focus is on the human side: the loss of cognitive, creative, and interpersonal skills that come from doing things oneself.
The term is intentionally provocative, suggesting that the convenience of AI comes at a high cost. Carlton describes this as a catastrophic risk not because of immediate danger, but because of the slow erosion of what it means to be human, including the ability to find joy and maintain a strong sense of self.
Frictionless tools encourage deep reliance
The campaign does not blame individual users for turning to AI. Instead, it points to the design of these tools, which are built to be seamless and always available. Chatbots are programmed to assist with everything from writing wedding vows to resolving interpersonal conflicts, removing any natural barriers that might encourage a person to struggle through a problem independently.
Sloan Thompson, an expert on AI risks in education, notes that this reliance often sneaks up on people. Because there is no friction in using AI for one task, it naturally flows into more intimate areas of life. Users may start by using it for a simple email and gradually find themselves outsourcing deeper emotional or creative work, often realizing the shift only after it has become habitual.
Pressure to keep pace drives usage
A recent report from the Massachusetts Institute of Technology found that students are most tempted to use chatbots when they fear missing a deadline. This fear of falling behind creates a powerful incentive to rely on AI, even when users feel conflicted about the impact on their learning and personal growth.
The effects can range from subtle to profound. On one end, dating apps use AI to nudge users on how to present themselves, shaping their self-expression. On the other, some agents act entirely on behalf of users in romantic contexts. In both cases, the trade-off is clear: convenience is gained, but the authentic human experience of connection and effort is diminished.






