Thinkorithm

AI and Society: Trust, Information and Fairness

How international principles and new law approach the social questions raised by AI — from fairness and transparency to the integrity of information.

Published · 2 min read

As AI systems are built into everyday services, the questions around them become social as well as technical. Who is responsible when a system is wrong? How do people know what is real? Is everyone treated fairly? Governments and international bodies have begun to answer in broadly similar terms.

Shared principles

In May 2019 the Organisation for Economic Co-operation and Development adopted the OECD AI Principles, the first intergovernmental standard on AI. They were updated in May 2024 to reflect technological and policy developments, including generative AI, and now address privacy, intellectual property, safety and information integrity more directly.

In November 2021 UNESCO’s 193 Member States adopted the Recommendation on the Ethics of Artificial Intelligence, the first global standard on AI ethics. It rests on values such as human rights and dignity, diversity and inclusiveness, and environmental flourishing, and turns them into policy recommendations in areas from data governance to gender equality.

Trust and transparency

A recurring idea in these frameworks is that people should know when they are dealing with an AI system and should be able to understand, at an appropriate level, how decisions affecting them are made. The European Union’s AI Act turns parts of this into law: it includes transparency obligations, such as informing people that they are interacting with an AI system and marking AI-generated content.

The integrity of information

Generative AI makes it cheap to produce realistic text, images, audio and video. That is useful for creative and practical work, but it also lowers the cost of misleading content. NIST lists “information integrity” among the risks that generative AI creates or makes worse, alongside confabulation and harmful bias.

Fairness

Systems trained on historical data can reproduce patterns of unequal treatment. Frameworks such as NIST’s AI Risk Management Framework do not promise to eliminate bias; they ask organisations to identify, measure and manage it throughout a system’s life, and to keep humans accountable for outcomes.

Sources

  1. OECD. OECD updates AI Principles to stay abreast of rapid technological developments, 2024; Recommendation of the Council on Artificial Intelligence.
  2. UNESCO. Recommendation on the Ethics of Artificial Intelligence, 2021.
  3. Regulation (EU) 2024/1689 (Artificial Intelligence Act). EUR-Lex.
  4. National Institute of Standards and Technology. Generative Artificial Intelligence Profile, NIST AI 600-1, 2024.
  5. National Institute of Standards and Technology. AI Risk Management Framework (AI RMF 1.0), NIST AI 100-1, 2023.