Remote AI Safety Analyst – Flexible, Adversarial Testing

About The Role

What if your critical thinking and curiosity could make AI safer for millions of people? We’re looking for AI Safety Analysts to put cutting‑edge AI systems under pressure — probing for harmful outputs, unexpected behavior, and safety blind spots that need to be caught before they reach the real world.

• Organization: Alignerr
• Type: Hourly Contract
• Location: Remote
• Commitment: 10–40 hours/week

What You’ll Do

• Probe AI systems with challenging, adversarial, and edge‑case inputs to surface unsafe or harmful outputs
• Identify responses that are inappropriate, dangerous, misleading, or otherwise problematic
• Document safety issues with precise, well‑structured examples and explanations
• Rate AI responses using safety and helpfulness evaluation frameworks
• Follow structured red‑teaming protocols and testing guides
• Work independently and asynchronously on your own schedule

Who You Are

• A natural critical thinker who enjoys finding gaps, inconsistencies, and edge cases
• Comfortable exploring unusual, sensitive, or unexpected scenarios in a structured way
• Strong written communicator — you can describe problems clearly and precisely
• Genuinely interested in AI safety, responsible technology, or the ethics of AI systems
• Self‑motivated and reliable when working independently
• No cybersecurity, AI, or technical background required

Nice to Have

• Experience in research, journalism, policy, ethics, or quality assurance
• Familiarity with AI chatbots or language models as an end user
• Background in logic, philosophy, or analytical writing

Why Join Us

• Work on real AI safety projects alongside leading research labs
• Fully remote and flexible — work when and where it suits you
• Freelance autonomy with the structure of meaningful, task‑based work
• Contribute to work that genuinely matters — helping ensure AI is safe before it reaches the public
• Potential for ongoing work and contract extension as new projects launch

#J-18808-Ljbffr…