Anthropic·Product Manager·Onsite - Product Sense / Strategy·Senior

Senior

Jun 2026

Summary

Interviewed for a PM role at Anthropic and got hit with a question about anthropomorphizing LLMs that I did not see coming from that angle.

Questions Asked (1)

What are the risks of assuming that large language models think or feel the way humans do?

Product StrategyAdaptability & AmbiguityProduct Sense & Ideation

Author's notes

I fumbled the opener because I started listing obvious UX stuff like users over-trusting the model, but they kept pushing and I realized pretty late they wanted something more fundamental.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Acknowledge the natural human tendency to anthropomorphize LLMs, then systematically unpack the risks across product, ethical, and strategic dimensions. Emphasize that while LLMs can simulate understanding, conflating simulation with genuine cognition leads to flawed product decisions and user harm. Conclude by proposing safeguards like clear communication, user education, and robust evaluation methods.

Pro tip: Frame the risks in terms of product outcomes—such as user trust, safety, and feature design—rather than abstract philosophy, to demonstrate business acumen. Show awareness of Anthropic's safety-first mission by highlighting how anthropomorphism can undermine responsible AI deployment.

1. Define the anthropomorphism trap

Explain that LLMs are statistical models trained on human text, not sentient beings, and that assuming human-like thought or feelings is a cognitive bias. This sets the stage for analyzing concrete risks.

2. Identify product risks

Discuss how anthropomorphism can lead to overpromising capabilities (e.g., emotional support bots), misaligned user expectations, and features that exploit emotional attachment. This can result in user disappointment, ethical breaches, and reputational damage.

3. Explore ethical and safety risks

Highlight dangers like users forming unhealthy dependencies, privacy violations from oversharing, and reinforcement of harmful stereotypes. Also note the risk of dismissing real AI harms (e.g., bias) by focusing on imagined sentience.

4. Consider strategic and competitive risks

Explain that anthropomorphic framing can lead to misguided investments in 'human-like' features at the expense of core capabilities, and can attract regulatory scrutiny. Competitors may exploit anthropomorphism irresponsibly, setting back industry trust.

5. Propose mitigation strategies

Suggest product design principles: transparent communication about AI limitations, user education on how LLMs work, and evaluation metrics that prioritize truthfulness over human-likeness. Advocate for cross-functional collaboration to uphold ethical standards.

Key Points to Mention

Anthropomorphism can lead to over-trust and over-reliance on LLM outputs, causing users to accept incorrect or harmful information.
Designing for emotional attachment may create ethical issues, especially for vulnerable users, and could violate privacy if users share sensitive data.
Product features that imply sentience (e.g., 'the AI cares about you') can backfire when the model fails to meet expectations, damaging brand trust.
Focusing on human-like qualities may divert resources from improving accuracy, safety, and robustness—core aspects of product value.
Regulatory and legal risks: anthropomorphic marketing could be deemed deceptive, leading to fines or restrictions.
Mitigation includes clear disclaimers, user education, and designing interactions that encourage critical thinking rather than blind trust.

AI-generated suggestions, not part of the candidate's original notes. May be inaccurate — verify before relying on them.