Font size:
Print
Data Annotations
Context:
Thousands of gig workers are becoming the backbone for training artificial intelligence-based large language models (LLMs) by taking on microtasks.
More on News
- Microtasks include transcribing audio, labeling images, translating text, and marking objects in self-driving clips and chatbot responses.
- India is becoming a major hub for data annotation services with a diverse workforce contributing to creating top-notch datasets.
Key Highlights:
- According to NASSCOM the global data annotation market is projected to reach $10 billion by 2028 growing at a rate of 27.2% annually.
- India’s data annotation market is expected to exceed $7 billion by 2030.
About Data Annotation
Also known as Data labelling is a critical step in creating high-quality datasets to train AI models.
- It ensures accuracy, preventing errors, and establishing safeguards against inappropriate or harmful content.
- Manually annotating data by human annotators or leveraging machine learning algorithms facilitates this process.
Annotation-as-a-Service
- It refers to a system or platform that allows the storage, management, and retrieval of metadata or additional information.
- Including tags, temporal data, spatial data, or other types of metadata.
- It is important for supporting a wide range of applications in different domains.
- It is experiencing significant growth in popularity, especially in India, as noted by the author and CEO of AI startup ScreeAI.
