Trust & Safety - English speaking

Written by the T&S Network from Accenture's posting.

Content moderation and responsible AI work for platform clients, covering user profiles, audio, video and text. The team acts as a line of defense for online content quality and safety, and also tests and challenges AI systems and platforms, including generative AI.

What you'd do

• Review, classify, or remove content according to client guidelines using specified tools and channels.
• Investigate, escalate, and resolve complex content issues and relay them to the wider Trust and Safety team.
• Design prompts that test large language models on sensitive subjects such as medical and legal advice.
• Assess model responses for appropriateness, refusals, and appended disclaimers against client definitions.
• Collaborate with team members to build a shared understanding of responsible AI issues arising from model responses.
• Track changing client policies and industry developments in content moderation and responsible AI regulation.
• Help draft best practices and rules for content moderation and responsible AI governance.

What it asks for

• Must handle emotional strain and stress well, with real coping ability.
• Must read and write English well and understand material quickly.
• Ability to perform duties objectively, without personal bias, in line with client policies.
• Strong attention to detail.
• Comfort synthesizing information from multiple streams.
• Sound reasoning and the ability to make judgment calls under pressure.
• Deep familiarity with the internet, platforms, and internet culture (preferred)

Worth knowing

• Exposure to objectionable or disturbing content, including graphic, violent, sexual, or egregious material, and adversarial scenarios for AI systems.
• Participation in a well-being support program designed for the Trust & Safety community is required.
• Shift work.

Our best summary of the posting, and it can be wrong. The employer's posting is what counts: read it in full on Accenturenorth_east before you apply. If we got something wrong, report this listing.