Skip to content
AIPIA — Italian AI Professionals Association
Code of Ethics

Six principles every AIPIA member commits to

The AIPIA Code of Ethics translates international AI frameworks — Rome Call, UNESCO, OECD, Council of Europe, EU AI Act — into binding commitments that govern day-to-day professional practice. Membership requires acceptance of the Code; breach is grounds for sanction.

  1. 01

    Competence

    Members maintain demonstrable, current expertise in their area of practice. They acknowledge the boundaries of their knowledge and seek peer or expert review when working outside familiar territory. Continuous learning is treated as a professional duty, not an optional extra.

  2. 02

    Transparency

    Members document AI systems in a way that allows informed scrutiny by clients, employers, regulators, and — where applicable — affected individuals. Capabilities and limitations are communicated honestly. Black-box deployments without explanation are avoided wherever decisions affect rights, opportunities, or safety.

  3. 03

    Fairness

    Members actively assess AI systems for discriminatory impact across protected attributes and contextually relevant groups. They design data collection, model training, and evaluation pipelines to detect and mitigate bias. When fairness cannot be reasonably ensured, the system is not deployed in contexts where it would cause material harm.

  4. 04

    Accountability

    Members accept responsibility for the AI systems they design, deploy, advise on, and approve. They keep records sufficient to reconstruct decisions, retain logs of model behaviour where appropriate, and cooperate with internal audits and external regulators. Responsibility is not delegated to "the algorithm".

  5. 05

    Privacy and data protection

    Members handle personal and sensitive data in accordance with applicable law — GDPR within the EU, sectoral rules elsewhere — and the higher standard where multiple frameworks apply. Privacy by design is treated as default. Members refuse to participate in mass surveillance, untargeted profiling, or covert data harvesting.

  6. 06

    Societal benefit

    Members consider the broader social, environmental, and human consequences of the AI systems they build. They prioritise applications that expand human capability over those that diminish it. They decline projects that would cause foreseeable harm disproportionate to any reasonable benefit, regardless of remuneration.

Alignment

Built on international AI ethics frameworks

The Code does not stand alone. It anchors to the major instruments shaping AI governance worldwide, so that members operating across jurisdictions can rely on a single coherent professional standard.

  • Rome Call for AI Ethics (2020)

    AIPIA is an Official Endorser. The Code adopts the Call's principles on transparency, inclusion, responsibility, impartiality, reliability, and security and privacy.

  • UNESCO Recommendation on AI Ethics (2021)

    The first global instrument on AI ethics, adopted by 193 member states. The Code reflects its values of human rights, environmental flourishing, diversity, and proportionality.

  • OECD AI Principles (2019, updated 2024)

    Adopted by 47 governments. The Code aligns with inclusive growth, human-centred values, transparency, robustness, and accountability.

  • Council of Europe Framework Convention on AI (2024)

    The first binding international treaty on AI. The Code anticipates its operational requirements on human rights, democracy, and the rule of law.

  • EU AI Act (Regulation 2024/1689)

    The Code is designed to operate within and complement the regulation. Members deploying high-risk systems comply with both.

Enforcement

The Code is enforced, not just declared

Self-assessment

Members are expected to audit their own practice against the Code at least annually and act on findings.

Peer review

AIPIA facilitates confidential peer consultation for members navigating difficult ethical decisions.

Complaints procedure

Third parties can file complaints against members for breaches. The Ethics Committee reviews each case in adversarial procedure with right of reply.

Sanctions

Outcomes range from formal warning to suspension or expulsion from the association, depending on severity. Decisions are appealable to the Members' Assembly.

Practice AI responsibly

Join AIPIA, accept the Code, and gain access to peer support, training, working groups, and EU-recognised credentials.

Become a Member Rome Call for AI Ethics