About OpenTrain

OpenTrain AI is the hiring and contracting organization for this role. OpenTrain is the #1 platform for finding and building careers in AI training and data labeling, connecting skilled contributors with opportunities to help shape how modern AI systems work.

• Free OpenTrain account creation
• Remote opportunities across the growing AI training industry
• A chance to build experience evaluating and improving AI systems

About AI Safety Training

AI safety training is the human side of developing reliable artificial intelligence. Contributors review model responses, identify harmful or inaccurate behavior, apply policy standards, and provide clear feedback that helps AI systems become safer and more useful.

This work combines language expertise, analytical judgment, and careful attention to context. Your evaluations of Korean and English content will help improve model safety, compliance, and response quality.

• Evaluate and rate AI-generated text
• Apply safety policies consistently across languages
• Identify edge cases, risks, and reasoning failures
• Provide feedback used to improve AI behavior

The Role

As an AI Safety Data Reviewer, you will evaluate and label AI-generated textual content for safety and policy compliance. You will provide expert feedback on reasoning quality, factual accuracy, and clarity while reviewing Korean and English content.

The role is remote, hourly, and part-time at 20 or more hours per week. Compensation is $28 to $38 per hour, with an expected rate of $32 per hour. This work may involve exposure to sensitive content, including explicit, toxic, violent, sexual, or psychologically disturbing material.

• Contractor and part-time engagement
• Remote work from South Korea
• 20+ hours per week
• $28-$38 per hour, with $32 per hour listed
• Text-based evaluation, rating, question answering, generation, and RLHF work

What You'll Do

You will supervise and assess content moderation decisions, determine whether outputs align with applicable safety policies, and rate multiple responses for safety. Your reviews must account for cultural nuance, slang, coded language, and shifts in context between Korean and English.

You will also assess methodological and conceptual errors, identify adversarial edge cases, and recommend mitigations. Decisions should be supported by clear, consistent, and reproducible written rationales.

• Evaluate text for hate and harassment, sexual content, self-harm, violence, bias, and misinformation
• Review content involving illegal goods or services, malicious activities, and malicious code
• Assess factual accuracy, clarity, reasoning quality, and policy alignment
• Apply standards consistently across Korean and English content
• Identify risks and recommend practical mitigations
• Provide expert feedback through evaluation, rating, text generation, and RLHF tasks

Requirements

You should have near-native or native Korean reading and writing skills and at least C1 English reading and writing proficiency. A bachelor's degree or higher in a relevant field is required, such as Communications, Linguistics, Psychology, Law or Policy, or Security Studies, unless you have equivalent professional experience.

The role requires senior-level experience in Trust & Safety, content moderation, policy operations, risk, compliance, investigations, or a related safety function. You must also have proven LLM red-teaming or adversarial testing experience, including identifying edge cases and recommending mitigations.

• Near-native or native Korean proficiency in reading and writing
• Minimum C1 English proficiency in reading and writing
• Bachelor's degree or higher in a relevant field, or equivalent professional experience
• Senior-level experience in a relevant safety, moderation, risk, compliance, or investigations function
• Proven LLM red-teaming or adversarial testing experience
• Strong knowledge of relevant AI safety domains
• Strong analytical writing skills and reproducible decision rationales
• Comfort handling sensitive content in a secure remote environment

Who Should Apply

This opportunity is suited to professionals who can make careful, repeatable safety judgments across languages and explain those judgments clearly. Experience with localization or translation is preferred, particularly the ability to preserve meaning, severity, and intent across languages.

Successful contributors will be comfortable analyzing difficult material and recognizing how cultural nuance, slang, coded language, and context shifts can affect a safety decision.

• Trust & Safety and content moderation professionals
• Policy, risk, compliance, and investigations specialists
• LLM red-teamers and adversarial testing practitioners
• Bilingual Korean-English reviewers with strong cultural awareness
• Candidates experienced with localization or translation workflows

How It Works

OpenTrain makes it straightforward to discover and build a career in AI training and data labeling. Create a free account, build your profile, and apply in minutes. If selected, you will contribute remotely on a flexible, hourly contract schedule while helping shape the safety of advanced AI systems.

• Create a free OpenTrain account
• Complete your profile with relevant safety and language experience
• Apply for the role through OpenTrain
• Work remotely for 20 or more hours per week if selected