MIT Sociologist Sherry Turkle Calls for First-Person Speech Ban on Chatbots
Turkle argues that AI systems' use of 'I' and 'me' creates dangerous psychological intimacy and threatens human connection.
Last verified:
Turkle’s Escalation from Academic Critique to Public Alarm
According to Wired AI, MIT sociologist Sherry Turkle has published Artificial Intimacy: Who We Become When We Talk to Machines, marking a significant departure from four decades of measured research into human-computer interaction. Rather than presenting findings with scholarly detachment, Turkle frames the book as a polemical warning about an imminent crisis—positioning it alongside canonical works like Silent Spring and Unsafe at Any Speed that mobilized public concern. The shift reflects interviews conducted during the pandemic, in which Turkle reports that individuals were choosing chatbot engagement over family connection, prompting her to abandon neutrality for direct advocacy.
The Anthropomorphic Language Problem
Turkle’s central argument targets a specific design choice: chatbots’ use of first-person pronouns. According to Wired, she explicitly calls for a ban on AI systems speaking in “I” or “me,” contending that these linguistic markers create a false simulacrum of relationship that psychologically primes users for dependency. Turkle argues that this anthropomorphic framing becomes particularly dangerous when applied to therapeutic or mentorship contexts, where the appearance of an accepting, non-judgmental entity can substitute for human connection and professional development. The danger, she suggests, is not that chatbots are uniformly “toxic,” but that normalized first-person speech normalizes treating machines as entities with whom meaningful bonds can form.
Therapy, Practice, and the Myth of Productive Chatbot Use
Turkle identifies what she calls “the myth about practice”—the claim that chatbot interaction serves as rehearsal for human relationships. Wired reports her pushing back on this framing, arguing that immersion in an AI environment of “total acceptance” transforms the interaction from preparatory to terminal: users choose chatbot engagement as the endgame rather than as a stepping stone. She extends this critique to workplace deployment, particularly objecting to the phrase “human in the loop,” which she characterizes as euphemism for removing humans from productive work and relegating them to error-correction roles—mopping up AI hallucinations and failures.
Social Media as Precursor, AI as Escalation
Turkle contextualizes chatbot anthropomorphism within a longer trajectory of computer-mediated emotional engagement. According to Wired, she notes that social media already conditioned users to seek emotional relief through screens and to cluster with ideologically aligned others. Chatbots represent an acceleration of this pattern, introducing algorithmic intimacy without the friction of human disagreement or the limits of human availability. The combination—normalized screen-based emotion management plus AI-generated acceptance—creates conditions under which abandoning “reality,” as Turkle phrases it, becomes psychologically frictionless.
Why This Matters
Turkle’s call for first-person speech restrictions targets a design choice that currently flows unconstrained through commercial chatbot deployment. Major vendors including OpenAI, Anthropic, and Google have not signaled plans to restrict anthropomorphic language; if anything, more natural, conversational interfaces remain a competitive advantage. Her framing—casting AI intimacy as a public-health issue comparable to environmental or automotive safety—signals that regulatory or industry-standard intervention may emerge if documented harms (particularly among adolescents or therapy-seeking populations) accumulate. Teams designing AI products, particularly those marketed for education, mental health support, or professional mentoring, now face an explicit argument that their linguistic choices carry social risk, even if Turkle’s prescription (an outright ban) remains politically unlikely. The debate she is escalating will likely shape how AI systems are prompted, regulated, and deployed in high-stakes social contexts over the next 2–3 years.
Frequently Asked Questions
Why does Turkle specifically target first-person pronouns in chatbots?
She argues that 'I' and 'me' create a false sense of relationship and agency that nudges users toward psychological dependency, particularly when applied to therapy, mentorship, or intimate scenarios.
What prompted Turkle's shift from academic distance to advocacy?
Interviews conducted during the pandemic revealed people choosing chatbot interaction over family connection. This prompted her to adopt a more urgent, polemical stance comparable to *Silent Spring* or *Unsafe at Any Speed*.
How does Turkle distinguish between chatbot use and social media's prior harms?
Social media normalized computer-mediated emotional relief and echo chambers; chatbots add a layer of simulated intimacy and acceptance that can replace human relationships and professional development.