Graphic Designer - AI Image Evaluation (Remote)
About the Company
Outlier is an AI training platform owned and operated by Scale AI. The platform connects experts with projects that help train and evaluate advanced AI models. Outlier Experts contribute to improving AI systems by generating training data, reviewing model outputs, and providing detailed feedback.
About the Role
Outlier is seeking detail-oriented English speakers to evaluate AI-generated images, videos, and audio. This remote freelance opportunity involves interacting with AI voice models, comparing generated media, identifying quality issues, and providing clear written explanations of evaluation decisions. Contributors can work from anywhere within the currently accepted countries and set their own schedules, with no minimum weekly hours. The work is project-based, and tasks may vary depending on expertise, skills assessments, location, and project requirements.
Key Responsibilities
Conduct live, scenario-based conversations in English with AI voice models and evaluate how natural, useful, and clear their responses sound Compare AI-generated images and videos against text prompts to determine which output follows the instructions more accurately Evaluate edited images and videos against their original versions Inspect generated media closely, including zooming in on images and replaying audio or video, to identify quality issues Identify visual problems such as distorted anatomy, AI-generated artifacts, text errors, flickering, and unnatural motion Evaluate audio quality, pronunciation, naturalness, and consistency between audio and video Rate outputs based on prompt adherence, quality, realism, and other evaluation criteria Write specific, evidence-based explanations in English describing observed issues and supporting evaluation decisions Label first-person videos showing people or robots performing hands-on tasks Break video clips into individual actions and create accurate captions by reviewing and correcting AI-generated drafts Record short voice explanations when required by specific evaluation tasks Common Project Tasks Image Evaluation: Compare AI-generated images created from the same text prompt and assess instruction adherence, visual quality, and visible AI artifacts. Some tasks involve comparing edited images with their original versions. Video Evaluation: Evaluate AI-generated videos against text prompts or compare edited videos with their source clips. Video Caption Labeling: Review first-person videos of people or robots performing practical tasks, divide clips into individual actions, and write precise captions based on the footage. Audio and Voice Evaluation: Assess AI-generated speech for naturalness, pronunciation, and similarity to a target voice. Some projects may require recording a short spoken explanation.
Required Qualifications
Fluent English communication skills, including strong written communication and natural conversational ability Strong attention to detail and patience when reviewing visual and audio content Ability to identify subtle errors by carefully inspecting, zooming in on, or replaying media Comfort with realistic, scenario-based conversations and thinking on your feet Experience writing concise, evidence-based explanations or reviews Ability to make clear evaluation decisions when both available options have flaws
Preferred Qualifications
Currently pursuing or holding a Bachelor's degree or higher Requirements to Sign Up Valid identification and a mobile phone from your country of residence An up-to-date resume highlighting relevant expertise A LinkedIn profile showing educational background and work experience Compensation and Payment Project-based earnings of up to USD $15 per hour Additional incentives through Missions, Outlier's surge-pricing program Contributors earn an average of 7.5% in additional earnings through Missions, while top-quartile contributors increase their earnings by an average of 11% Payments are processed weekly on Tuesdays for tasks completed during the previous pay period, which runs from Tuesday to Monday at midnight UTC Available payment methods include PayPal, AirTM, and ACH transfer Rates vary according to expertise, skills assessment, location, and project requirements Higher rates may be available for advanced qualifications, including PhD-level expertise Lower rates may apply to non-core work, including initial project onboarding and project overtime phases Certain projects may offer additional incentive payments, subject to their individual payment terms Additional incentive figures are based on payments made during the previous six months and are updated quarterly. Benefits and Work Flexibility Work remotely from anywhere within the currently accepted countries Set your own schedule with no minimum weekly hours Choose to work five, 40, or another number of hours per week based on availability Free access to leading large language models valued at $200 per month Opportunity to collaborate with experts across different fields Contribute to the evaluation and improvement of pre-release AI models
How to Apply
Create an account on the Outlier platform Upload your resume Submit valid identification Complete the required onboarding and project tasks to begin earning Eligible Countries Currently accepting applicants from Australia, Canada, Ireland, New Zealand, the United Kingdom, and the United States. Employment Type and Work Authorization 1099 freelance contract opportunity through the Outlier.ai platform Contributors join as independent contractors, not employees Internships, employment, and visa sponsorship are not offered Applicants must be authorized to work in their country of residence International students may be eligible to sign up depending on their visa status. They should consult a tax or immigration advisor regarding their individual circumstances.