Interpretation Risk & Misuse Lead

About Second Pass

Second Pass gives users one reading of a conversation transcript and then stops. Users cannot ask it to clarify or revise that reading, so a tentative output can still be over-trusted or projected onto.

This role owns that risk.

About the Role

We are hiring an Interpretation Risk & Misuse Lead to identify and reduce risks caused by how users read, project onto, or over-trust product outputs.

The focus is false closure, authority attribution, paranoid or confirmatory readings, and misuse caused by interpretation rather than system behavior.

This is an epistemic-risk role, separate from moderation or therapy.

Responsibilities

• Identify places where users may read outputs as truth, diagnosis, or intent.
• Stress-test example outputs, labels, and framing for projection and over-confidence risk.
• Review wording for how definitive it feels.
• Define when ambiguity or restraint is safer than clarity.
• Recommend removing or revising outputs that create false closure, even when they are accurate.
• Partner with the Product Integrity Lead to address interpretation risks without expanding scope.

Out of scope

Adding features, expanding interpretive depth, user reassurance, advice, engagement optimization, and resolving user problems or relationships are outside this role.

Who You Are

• Notice the gap between what a sentence says and what a user may infer.
• Comfortable removing outputs that feel useful but invite over-trust.
• Skeptical of definitive framing in ambiguous human contexts.
• Prefer restraint to adding more explanation.

Relevant Experience

Useful backgrounds include:

• Trust & Safety or AI policy (interpretation-focused)
• AI red-teaming or misuse research
• Linguistics, discourse analysis, or behavioral risk
• Editorial or review roles involving high-stakes language decisions

Formal AI engineering experience is not required.

Compensation

• Base Salary Range: $220,000–$320,000 USD (depending on experience and scope)
• Equity / Long-Term Incentives: Available and negotiable

Logistics

• Location: Remote (United States)
• Employment Type: Full-time
• Reporting: Direct to founder; works closely with Product Integrity Lead
• Visa Sponsorship: Not available at this time

How to Apply

Please submit:

1. An example of language you removed or softened because it was risky despite being accurate.
2. One interpretation-driven failure mode you believe AI tools underestimate.

Concise answers preferred.