AI Dictionary › AI Fundamentals

AI Bias

Bias (nei sistemi AI)

Bias in AI systems is a model's systematic tendency to produce distorted or discriminatory results toward certain groups, topics or perspectives. Bias doesn't come from the model's "bad intentions": it is absorbed from training data, which reflects the prejudices present in human text, and from design choices.

Definition

Concrete examples: a CV screening system that penalizes female names because it was trained on unbalanced hiring history; an image generator that associates certain professions with a single gender; a language model that treats identical requests differently depending on the language they're phrased in.

For professional use, especially in HR, credit, healthcare and justice, bias is a legal risk as well as an ethical one: the EU AI Act classifies these uses as "high risk" precisely for this reason. Mitigation: curated datasets, fairness testing, human oversight on decisions that matter.

Related terms

• Training Data
• EU AI Act
• AI Hallucination
• Explainability (XAI)
• Machine Learning

More in AI Fundamentals

• LLM (Large Language Model)
• Token
• Context Window
• Temperature
• AI Hallucination
• Fine-tuning
• RAG (Retrieval Augmented Generation)
• Embedding

Put it into practice

• Real scenario: apply this concept with an instant AI verdict.
• Daily gym: short drills to train dictionary concepts.
• Playoff game: challenge other players, elimination-style.

📱 Download the Android app (beta) iOS coming soon

Say what you mean. Get what you need.

Grace Certified, the AI coach that trains and certifies your prompt engineering, by Agora Intelligence.