As the Digital Forensics Conference approaches us, taking place in Sava Centar on November 12th & 13^th, we had the opportunity to talk to Matija Franklin, AI Safety Researcher (Ex OpenAI & Google DeepMind), about his motivations, career milestones and his research on AI manipulation becoming a crucial part of AI safety policies coming straight from the center of Europe, in Brussels

What was your motivation for getting into this line of work?

AI safety offers a unique opportunity for researchers who are passionate about understanding human behavior while striving to align AI systems with human values. The chance to contribute to this alignment is what drew me to the field.

What was the biggest challenge for you while working in this field, and what would you pick as the most important milestone of your career?

Staying current with the rapid pace of new research is a constant challenge. AI is advancing at an extraordinary rate, and keeping up with all developments can be overwhelming. As for the most significant milestone, it’s likely my work on AI manipulation and general-purpose AI systems, which influenced the EU AI Act’s approach to safety.

How did your research on AI manipulation and General-Purpose AI systems end up being incorporated into the EU AI Act? How does this act regulate the safe use of AI in Europe?

I collaborated with the Future of Life Institute and the Future of Humanity Institute, both of which play key roles in shaping policy in Brussels. Their influence helped elevate research insights to policymakers. As the Act is still evolving, we’re in the process of drafting Codes of Practice, and its real-world impact on safe AI use will become clearer as these guidelines are implemented.

How do you see the future of AI? To what degree do you see it being incorporated into our everyday lives?

In recent years, we’ve witnessed a surge in AI capabilities. The next phase is likely to be a widespread adoption of AI, seamlessly integrating into every aspect of daily life. I find this shift exciting, as AI will become so ubiquitous that it will fade into the background of our everyday experiences.