Project
The Goal
The aim of the InfoAICert project is to develop the first made-in-Italy platform for conformity assessment of high-risk AI systems, by leveraging on the world-class research leadership of University of Genoa’s sAIfer Lab's on AI security and trustworthy AI, and on the international leadership of the company Tinexta Infocert as trust service provider.
Artificial intelligence-based technologies are becoming ubiquitous, with their applications and performance growing at an unprecedented pace, while the expected socioeconomic benefits are enormous.
The attention of the scientific community and stakeholders has long been focused on the beneficial uses of AI. Less attention has been devoted to potential misuse, harmful uses, and the lack of robustness of AI that could compromise the safety of individuals and society.
In June 2023, the European Parliament voted in plenary on the proposal for the European Artificial Intelligence Act (AI Act), with the objective of preserving the benefits of AI-based technologies while reducing the risks to European citizens.
The AI Act identifies high-risk AI systems that may access the EU market only after undergoing a “conformity assessment” (a sort of CE marking for AI) to validate and verify the safety and reliability of the AI system. For example, Annex III of the EU AI Act classifies systems for the identification of natural persons as high-risk systems.
Providers and manufacturers of high-risk AI systems will be required to provide conformity assessment bodies with all the necessary documentation to access the EU market. Companies placing AI systems that are not classified as high-risk on the market will nevertheless be encouraged to undergo conformity assessment, as this represents a good practice clearly recommended by the EU AI Act. This will create enormous market demand, given that 77% of companies worldwide are considering the use of artificial intelligence systems.
Consequently, among other things, the EU AI Act requires transparency, traceability, and risk management for many AI systems, especially high-risk ones. Only AI companies that will meet these requirements, will be compliant and, consequently more transparent and competitive.
Unfortunately, the EU AI Act provides regulatory requirements but does not offer practical guidelines or technical tools for conducting conformity assessments in accordance with the legislation.
The vision of InfoAICert is that market demand for innovative products and services for the conformity assessment of AI systems will surge after 2026. To take advantage of this future market demand, it will be necessary to advance fundamental research and carry out a substantial amount of new industrial research over the coming years.
The InfoAICert project aims to develop the first Italian-made platform for the conformity assessment of high-risk AI systems, providing innovative services to AI providers and manufacturers seeking access to the EU market. The platform will strengthen the competitiveness of Italian industry by enabling conformity assessment services for both existing and new EU customers.
The platform will strengthen the competitiveness of Italian industry by enabling conformity assessment services for both existing and new EU customers.
InfoAICert will fill a current market gap by providing a solution to make AI systems:
Competitive
Compliant
Certified
The challenge
How can we build intelligent systems that are reliable, safe, and aligned with human values?
How can we ensure that an AI system does what we are asking for, and not something that is dangerous or wrong?
As also reaffirmed by the social needs and by the international laws, it is becoming increasingly necessary to:
Prevent AI from making harmful decisions
Ensure it follows human rules
Prevent malicious or uncontrolled use
Reduce errors in critical contexts
During 5 years (2025-2030), by combining scientific research, industrial expertise, and regulatory insight, InfoAICert will contribute to the creation of a European certification model for AI, similar to the CE marking used for physical products.
The project supports the EU’s vision of trustworthy, human-centric AI, strengthening Europe’s leadership in the development of safe, transparent, and accountable AI systems.
The infoAICert platform evaluation will cover: Security and robustness, Transparency and explainability, Fairness and non-discrimination, Privacy and data governance.
Made in Italy
Research driven
Industrial Expertise
Regulatory compliance
Contact us
Need more information? InfoAICert is constantly looking for new research and development partners, and for end-users to test the platform in a mature stage. We’re here to answer any questions you may have. Send us a message, addressed to Prof. Fabio Roli - Principal Investigator, and we’ll get back to you as soon as possible.
