Fairness

TECHNICAL DEFINITION

Fairness in AI refers to the principle and practice of designing and deploying artificial intelligence systems to produce equitable outcomes, avoiding discriminatory impacts on individuals or groups based on sensitive attributes like race, gender, or socioeconomic status.

BACKGROUND

A large language model (LLM) is a neural network trained on a vast amount of text for natural language processing tasks, especially language generation. LLMs can typically generate, summarize, translate and analyze text in many contexts, and are a foundational technology behind modern chatbots. Biased or inaccurate training data can make an LLM's output less reliable.

SYNONYMS & ALIASES

• Equity
• non-discrimination
• impartiality
• justice

USAGE NOTE

Achieving fairness in AI often requires careful data collection and bias mitigation strategies.

DEVELOPERS

Organizations developing technology related to Fairness.

• Google AI

  Develops research, tools, and guidelines to promote fairness, accountability, and transparency in AI systems, impacting AI engineering and prompt design practices.

• IBM Research

  Conducts research and develops open-source toolkits like AI Fairness 360 (AIF360) to help engineers identify and mitigate bias in AI models, focusing on fairness in AI development.

• Microsoft Responsible AI

  Provides responsible AI tools and services, including fairness assessment dashboards and guidelines, to help developers build and deploy AI systems that are equitable and just.

• Meta AI

  Engages in extensive research on fairness, bias detection, and mitigation in AI systems, often publishing findings and open-sourcing tools relevant to AI engineering practices.

• Hugging Face

  Offers an open-source platform and community that supports the development and deployment of AI models, emphasizing responsible AI practices, including fairness considerations in engineering and prompt usage.

• Salesforce AI Research

  Conducts research on ethical AI, including fairness and bias mitigation, with the aim of building more trustworthy and equitable AI systems for enterprise applications.

• Allen Institute for AI (AI2)

  A non-profit research institute dedicated to AI, often focusing on fundamental challenges like fairness, robustness, and interpretability in AI system design and evaluation.

• Pymetrics

  Develops AI-powered talent acquisition platforms designed to eliminate bias and ensure fairness in hiring decisions through scientifically validated assessments and algorithms.