Amped Software recently presented a roadmap at the European Parliament addressing AI’s impact on image and video analysis. We advocate for public awareness, stakeholder education, and regulations like the AI Act to ensure security and justice in the digital age.
Amped Software is committed to a fundamental objective: significantly reducing unsolved crimes and preventing wrongful convictions through rigorous and ethical application of digital multimedia evidence. While developing advanced software and delivering expert training are essential components of our work, we recognize the critical need for broader engagement.
In recent years, we have devoted more resources to raising awareness among non-technical stakeholders, including policymakers, judges, prosecutors, attorneys, and journalists. Our focus is on educating them about the inherent power and potential fragility of digital image and video evidence.
Furthermore, the rapid evolution of artificial intelligence presents both significant challenges and transformative opportunities within our field. Building upon our previous experiences at the European Parliament in 2022 and 2023, where we introduced our influential document “Video Evidence Principles for the Use of Video Evidence for Public Safety and Criminal Justice”, we returned this year with a new initiative. We presented a high-level roadmap outlining strategic approaches to address the evolving landscape of AI in image and video forensics. This roadmap emphasizes the importance of public awareness, comprehensive education for all relevant stakeholders, and the necessity of robust regulatory frameworks, such as the AI Act.
The roundtable, hosted by EU parliament members Hon. Francesco Torselli and Hon. Elena Donazzan, has seen impactful contributions from key players in our field, such as:
- Stephen Kavanagh, Secretary General International Centre for Missing & Exploited Children (ICMEC – Global)
- Giovanni Tessitore, Senior Technical Director of Italian State Police, Scientific Police Service
- Arianna Lepore, Coordinator, Global Initiative on Electronic Evidence, Terrorism Prevention Branch, Division for Treaty Affairs, UNODC
- Giuseppe Zuffanti, European Security and Defence College, Coordinator for Cyber issues and for the Cyber ETEE Platform – Project Manager/Training Manager for Cyber – National Expert
A Roadmap for The Use of Image and Video Evidence in the AI Era
Below is the roadmap we presented at the event: it summarizes the key steps needed to capitalize on opportunities and address challenges related to the impact of AI on image and video evidence in the coming years.
1. 📢 Raise Awareness
Photos and videos are the most common and powerful types of evidence. They could be key to solving a crime, but they are often mishandled or underutilized. Recognizing their immense value and applying forensic rigor is crucial for trust in investigations and justice.
2. 🔎 Ensure Proper Education and Tools
Forensic processing and analysis require properly collected evidence, trained professionals, and reliable tools. A rigorous and scientific forensic workflow allows for the faster solving of crimes and the reduction of wrongful convictions.
3. 🎭 Address AI’s Impact on Evidence
AI raises issues like deepfakes, manipulated photos, and false claims of AI interference. AI-generated images and videos can have an impact on misinformation and radicalization even when we know they aren’t real.
4. 🚨 Combat AI-Generated CSAM
AI-generated Child Sexual Abuse Material (CSAM) is a growing and often overlooked issue that wastes investigative resources, normalizes and commercializes exploitation, and often relies on training done on real images of children. Urgent action is needed.
5. ⚡ Leverage AI’s Potential
AI provides unprecedented capabilities: accelerating investigations, enabling previously impossible analyses, and supporting the well-being of forensic practitioners. Used responsibly, it may improve efficiency and accuracy.
6. ⚖️ Mitigate AI’s Risks
AI introduces challenges such as lack of transparency, inherent biases, and automation bias in human operators. AI should be used as a decision-support tool, not a substitute for expert judgment.
7. ✅ Follow and Contribute to Regulatory Safeguards
EU laws like the AI Act and GDPR help ensure that AI and sensitive data are used responsibly. Compliance enhances the safety of investigations, reduces the risk of miscarriages of justice, and upholds fundamental EU rights and values. Stay updated and contribute to regulatory improvements to balance innovation and security.
8. 🛡️ Commit to a Future of Security, Justice, and Democracy
The future of image and video forensics depends on our collective efforts. Policymakers, law enforcement, forensic experts, and technologists must work together to establish transparent, ethical, and effective solutions. By embracing innovation while maintaining rigorous standards as in other forensic disciplines, we can enhance security, strengthen justice, and uphold democratic values, ensuring that truth prevails in the AI era.
Conclusion
It’s been a few years since we embarked on this institutional engagement project. Focusing solely on training our users eventually reached its limits, as upper management, institutions, and judicial stakeholders struggled to grasp the fundamental challenges inherent in image and video evidence. Fortunately, we are already seeing initial positive outcomes from our efforts. With each passing day, more individuals are understanding these issues and taking them more seriously. If you have any ideas on how we can accelerate this transition, please don’t hesitate to reach out!
