The task involves intentionally trying to break AI systems by crafting clever prompts and scenarios to expose weaknesses and safety issues. Alignerr is looking for curious, analytical thinkers with strong English skills who can document failure modes systematically. No prior technical or cybersecurity background is required for this fully remote, flexible freelance role.