Daniel & Friends / Glossary / Model bias (AI bias)
AI and automationModel bias (AI bias)
Model bias is a systematic distortion in AI outputs that comes from unbalanced training data, model design or the way a task is framed. The model then favors or overlooks certain groups, topics or languages, for example when ranking CVs or generating images.
Daniel Votruba ·
In practice
Before launch, test the model on samples from your own market and language. For decisions about people, leave the final call to a human.
Related terms
- Explainable AI (XAI): Explainable AI (XAI) is the ability to show in understandable terms why a model reached a given result: which inputs mattered, how confident it is and what the answer rests on.
- AI governance: AI governance is the set of rules, roles and processes a company uses to manage its use of artificial intelligence.
- EU AI Act: The EU AI Act is the European regulation on artificial intelligence.
- Large language model (LLM): A large language model is a model trained on a huge amount of text that predicts the next words and can write, summarise and answer.
- Model evaluation (evals): Model evaluations are systematic tests that measure the quality of AI answers on a set of typical tasks with a verified correct result.
More terms in the group AI and automation
MCP (Model Context Protocol) · Multi-agent system · Machine learning (ML) · Multimodal model · Neural network · Jailbreak (AI) · NLP (natural language processing) · Inference and latency · No-code and low-code · Human in the loop · Open-weight model · Hallucination · Predictive analytics · Guardrails · Prompt · Grounding · Prompt engineering · Generative AI · Prompt injection · Function calling (tool use) · RAG (retrieval-augmented generation) · Foundation model · Reasoning model · Fine-tuning
Terms in practice
Need help putting them to work?
Terms are one thing, putting them to work is another. We go through which of them pay off for you and say so plainly. You get a first opinion for free and we reply within 24 hours.
Book a call