General principles for the use of AI at CERN
—
At CERN, artificial intelligence (AI) is seen in a variety of contexts. They are embedded in devices, software products, and cloud services procured by CERN, brought into the field by individuals, and developed in-house.
As endorsed by the CERN-wide AI Strategy, these general principles are designed to promote the responsible and ethical use, development, and deployment (collectively, “Uses”) of AI at CERN.
These are technology-neutral and apply to all AI technologies that become available.
This principle applies to all areas of CERN’s activities, including:
- AI for science and technology research: data analysis, anomaly detection, simulation, predictive maintenance and optimization of accelerator performance and detector operation, and
- AI for productivity and management applications: Document creation, note-taking, automatic translation, language correction and enhancement, coding assistance, workflow automation.
general principles
CERN, its personnel, and all persons using CERN computing facilities shall ensure that AI is used in accordance with the following principles:
- Transparency and explainability: Document and communicate when and how AI will be used and how it will contribute to specific tasks and decisions.
- responsibility and accountability: The use of AI, including its impact and resulting outcomes throughout its lifecycle, must not replace ultimate human responsibility and accountability.
- legality and conduct: Use of AI must be lawful, comply with CERN’s internal legal framework, and respect the rights of third parties and CERN’s Code of Conduct.
- Fairness, non-discrimination and “do no harm”: AI must be used in ways that promote equity and inclusion and prevent bias, discrimination, and all other forms of harm.
- Safety and security: To reduce the likelihood and impact of cybersecurity incidents, AI must be properly secured. AI must be used in a way that is secure, respects confidentiality, integrity, and availability requirements, and prevents negative consequences.
- sustainability: The use of AI must be evaluated with the goal of reducing environmental and social risks and increasing CERN’s positive impact on society and the environment.
- human surveillance: The use of AI must always be under human control. Its functionality and output must be consistently and critically evaluated and verified by humans.
- data privacy: AI must be used in a way that respects privacy and protection of personal data.
- non-military purpose: The use of AI at CERN must be for non-military purposes only.
