Periodic Reporting for period 1 - STARLIGHT (Sustainable Autonomy and Resilience for LEAs using AI against High priority Threats)
Okres sprawozdawczy: 2021-10-01 do 2022-09-30
The challenge
The growth of AI provides security opportunities and challenges for LEAs. The challenge is to grow LEA knowledge, skills and capabilities to capitalise on AI to deliver better and more efficient solutions to counter major security threats whilst also mitigating against criminal use of the same AI technologies. LEAs should achieve this whilst also complying with legal, ethical and organisational constraints needed to safeguard the fundamental rights and freedoms of EU Citizens.
Project Objectives
• Transform the potential benefits and risks of AI in support of security into an opportunity to contribute and boost the creation of a European approach to AI for LEAs.
• Contribute to the establishment of a strong EU AI-based security industry, thus enhancing the EU’s strategic autonomy in the field of AI for LEAs.
Concept and approach
STARLIGHT will build a sustainable European AI community of relevant stakeholders applying a co-creation and co-development approach to identify and create human-centric AI solutions supporting daily activities and addressing the major threats faced by LEAs. The project endeavours to provide early access to results to end-users facilitated by a streamlined and flexible methodology enabling rapid feedback, evaluation, and uptake of results.
Expected Results
Tools: Wide range of easy to integrate, interoperable tools to support annotation for AI datasets, AI solutions to enable LEA prevention, detection and investigation capabilities (TRL7-8).
Knowledge: Improve and enrich LEAs’ knowledge and awareness of how AI can reinforce their operational and cybersecurity capabilities with a robust application of societal, legal and ethical principles.
Community: A sustainable community of AI stakeholders that will ultimately promote the creation of a European AI Hub for LEAs.
Six Use Cases are addressed (broken down into 12 scenarios): Counter-terrorism, Child Sexual Exploitation, Border & External Security, Cybersecurity & Cyber threat intelligence, Addressing information overload in Serious Organised Crime (SOC) and Protection of public spaces. They are the result of a joint analysis of operational scenarios, needs and desired functionalities of the concerned stakeholders’ community.
A focus was made on identifying the main legal and ethical frameworks that are applicable to STARLIGHT, as well as societal concerns that may arise from its implementation. A comparative study was conducted on various national legal frameworks on data handling and access challenges that may have an impact on the data sharing within the STARLIGHT partners for R&D purposes.
A first action was to identify available data-oriented tools (collection, annotation, anonymisation and generation) and datasets and to evaluate their maturity. It represents 16 tools and 81 datasets of which 18 new ones created in STARLIGHT. Different data modalities (image, audio, text, video) are processed. Privacy-friendly data processing strategy is at the heart of the development. In addition, quality assessment is taken into account to identify possible biases in the distribution of the data or classes under-represented. The use of these tools was then prioritised according to their maturity and the functionality required for the use cases and scenarios selected.
In order to speed up the process, it was decided to run parallel co-creation and co-development cycles to tackle specific sub-scenarios. These 6-month co-development cycles are led conjointly by a LEA and a technical partner. They will give rise to hackathons for a first evaluation and to a final pilot demo in operational environment. A first set of 5 co-dev cycles was launched in October 2022 for a demo planned in March 2023. A second set of cycles and their LEA/technical partner tandems is already defined.
Understanding AI vulnerabilities is of utmost importance and preliminary work has been done to study the latest threat intelligence strategies and to establish a basic methodology for self-assessment of risks to anticipate, predict and investigate cyber-attacks.
Finally, the conceptualisation of the orchestration framework and connectors has been developed to facilitate the integration of STARLIGHT tools in operational pipelines through a unified approach. At this stage, mockups of the STARLIGHT repository are designed, and their implementation is undergoing. This will be enriched by the “lessons learned” from the first co-development cycles.
The important involvement and commitment of the LEAs in the project, since the first stages, will help the solutions’ evaluation, testing and final adoption by STARLIGHT’s end-users. At this stage, LEA partners are aware about state-of-the-art technological advancements in AI and about the legal and regulatory framework for the usage of AI for the security domain.
All activities within STARLIGHT are working towards ‘Protecting freedom and security of Europe and its citizens’. STARLIGHT will then have a positive impact on society by disrupting criminal activities, reducing risks to society from criminality and addressing concerns of EU citizens. That’s why STARLIGHT has also placed significant emphasis on the research carried out in societal, legal and ethical considerations to ensure a consistent understanding and approach. In addition, STARLIGHT is organizing multimodal awareness campaigns targeting citizens and civil society which will give the project the opportunity to directly engage with a wider audience and demonstrate the positive impact the results of STARLIGHT can have on society.