Task Types
Explore the different types of AI training tasks you can contribute to.
Data Annotation & Labelling
Essential foundation work for AI training involving precise data categorization and identification.
Task Type 1Key Activities:
- • Text classification: Tagging sentiment, intent, or topic
- • Named entity recognition (NER): Identifying people, places, organizations in text
- • Image/video labeling: Bounding boxes, segmentation, object identification
- • Audio transcription: Converting speech to text, often with speaker or emotion tags
- • Multimodal alignment: Matching text descriptions to images or videos
Prompt Evaluation & Response Rating
Critical quality assurance work ensuring AI responses meet high standards of accuracy and safety.
Task Type 2Key Activities:
- • Ranking responses: Choosing the best among multiple generated replies
- • Scoring relevance, coherence, and factuality
- • Flagging harmful, biased, or misleading content
- • Comparing human-written vs. AI-generated text
Synthetic Data Generation
Creative content creation to expand AI training datasets with diverse, domain-specific examples.
Task Type 3Key Activities:
- • Writing prompts and completions for specific domains (e.g., medical, legal)
- • Generating diverse examples to improve model generalization
- • Creating adversarial examples to test model robustness
Search Relevance & Query Evaluation
Improving search and information retrieval systems through careful evaluation and optimization.
Task Type 4Key Activities:
- • Judging relevance of search results to user queries
- • Annotating query intent and user satisfaction
- • Evaluating autocomplete suggestions
Conversational AI Testing
Enhancing chatbot and conversational AI systems through comprehensive testing and feedback.
Task Type 5Key Activities:
- • Roleplaying user scenarios to test chatbot performance
- • Logging bugs or unnatural responses
- • Providing feedback on tone, empathy, and engagement
Bias & Fairness Audits
Ensuring AI systems are fair and inclusive by identifying and addressing potential biases.
Task Type 6Key Activities:
- • Flagging biased outputs across gender, race, religion, etc.
- • Testing prompts for differential treatment
- • Suggesting inclusive alternatives
Language Expansion & Localization
Making AI accessible globally by expanding language support and cultural understanding.
Task Type 7Key Activities:
- • Translating prompts and responses
- • Annotating dialects, idioms, and cultural nuances
- • Validating model performance in low-resource languages