Digital Social Sciences (2021 Cohort)
The literature on human control of automated systems has a long, multidisciplinary history of research seeking to address the question of how to reasonably integrate ever more capable and complex systems in various areas of human activity. Striking a balance between the opportunities that such systems bring to human endeavours and their associated operational risks, such as the loss of human operators’ situational awareness leading to harmful consequences. This PhD thesis explores this fragile balance in the current applications of AI in Computer Network Operations (CNOs). In doing so, it seeks to test the utility of existing human control frameworks in under-studied, high-stakes operational environments and to propose an approach to human control of AI based on the human-machine teaming paradigm.
Andreas did his BA at King’s College London in the department of War Studies and wrote a thesis on China’s People's Liberation Army’s Strategic Support Force and its cyber strategy. He then did his MSc at the University of Oxford’s Oxford Internet Institute, where he researched and demonstrated the useful of Large Language Model’s in conducting sybil attacks on democratic government’s public consultation websites and fora.
He is currently developing the Algorithmic Resistance Cookbook, a guide to using data-driven tools and techniques to practice resistance against intrusive and repressive aspects of present-day algorithmic culture.
Andreas is also building Fileverse.io, an open source and peer to peer file sharing and collaboration tool designed to be privacy-preserving and censorship resistant. Fileverse is supported by over 14 000 donors.
He has published a number of published articles, including in prestigious journals like Nature (Machine Intelligence) and Cell (One Earth): https://scholar.google.com/citations?user=ExvifmMAAAAJ&hl=en
And with research institutions such as the Alan Turing Institute: https://cetas.turing.ac.uk/publications/artificial-intelligence-national-security-predictability-problem
Prizes and Fellowships:
- Awarded a grant by the Justice, Equity and Technology Project at LSE
- Awarded a grant by the Gitcoin DAO
- Fellowship on AI in security by Google Europe
- Fellowship on the climate crisis and AI by the Vodafone Institute for Society and Communication
- Awarded a grant by the Minderoo-Oxford Challenge Fund in AI Governance
- Awarded a scholarship by Mansfield College
- Dean’s Award a prize by King’s College London