Remote AI Model Quality Evaluator

Outlier (by Scale AI); United States; Remote; USD 17,000 - 24,000; Full time

Benefits offered by this job

Flexible schedule

Free access to leading LLMs

Join 900K+ Outlier Experts

Job summary

Outlier, owned and operated by Scale AI, seeks detail-oriented English speakers to review AI-generated media. You will roleplay live conversations with AI voice models, evaluate image/video prompts, and explain in writing what looked or sounded wrong, all remotely and on your own schedule.

Responsibilities include ranking naturalness, comparing media, inspecting outputs for issues, and writing concise explanations in English that point to observed errors.

Qualifications

• Fluent English with clear written communication and ability to explain observations concisely.
• High attention to detail and ability to notice inconsistencies in media outputs.
• Experience writing short, evidence-based explanations or reviews.
• Ability to make clear decisions when both options have flaws.
• Bachelor's degree or higher is a plus.

Responsibilities

• Hold live, scenario-based conversations in English with AI voice models and rate naturalness and clarity.
• Compare pairs of AI-generated images/videos and judge which adheres to the prompt.
• Inspect outputs closely to spot issues like distorted anatomy, flickering, or audio-video misalignment.
• Provide written explanations detailing what you saw or heard, with specific references.

Skills

Fluent English

Attention to detail

Clear written communication

Decision making

Bachelor's degree (nice to have)

Education

Bachelor's degree or above

Job description

Outlier, owned and operated by Scale AI, seeks detail-oriented English speakers to review AI-generated media. You will roleplay live conversations with AI voice models, evaluate image/video prompts, and explain in writing what looked or sounded wrong, all remotely and on your own schedule.

Responsibilities include ranking naturalness, comparing media, inspecting outputs for issues, and writing concise explanations in English that point to observed errors.