This project will research how a new constitutional right to "AI integrity" (protecting users' AI agents from tampering and/or ensuring their loyalty to their principals) could be established to prevent human disempowerment as people increasingly act through company-controlled AI intermediaries in markets and democratic participation.
About the project
I would like to research how the right to one's AI agent not being tampered with, and being built in such a way that it is loyal to its user/principal, could be enshrined constitutionally. I believe we are slowly drifting towards a world where people act, in the market or by voting, through an AI agent. Everything seems to point towards this being done in an intermediated way, such that AI companies like OpenAI might control the hardware and software of consumers' decision-making. This could lead to dystopian outcomes of human disempowerment. I believe that a day will come where it will be necessary to protect people's rights with regards to the integrity and loyalty of those agents. Consumer protection is insufficient, and ontologically inadequate.
See more details in the project proposal: https://docs.google.com/document/d/1yElckIazWnXmfKo8OLeWA6C5KjeSo3C0LXVFLJ_gHtU/edit?usp=sharing
Theory of change
By establishing AI agent integrity as a recognized category within fundamental rights discourse and providing actionable blueprints for legal enshrinement, we contribute to providing the AI safety community with a new governance tool that addresses the principal-agent alignment problem at the societal level; ensuring that even if technical alignment succeeds, the benefits accrue to users rather than being captured by intermediary AI companies.
Current AI safety efforts focus predominantly on aligning AI to human values but neglect the institutional layer where control over “aligned” AI systems could still be exploited; constitutional rights provide durable, hard-to-reverse protections that consumer law cannot; and early conceptual groundwork that shapes societal expectations is essential before AI-mediated agency becomes so entrenched that meaningful reform becomes politically infeasible.
Your role
Mentees should organize independently and collaborate actively, particularly in the initial exploratory phase of the project. They will be supported by a weekly meeting with the mentor, and receive asynchronous feedback on their work, as well as personal mentorship, where useful. Mentor involvement may vary around project milestones.
Prerequisites
Mentees collaborating on this project should have complementary backgrounds.
Law: At least one of the mentee will have a legal background. Knowledge of the preferred.
Other backgrounds are welcome (philosophy, sociology, policy, human-AI collaboration…) as well as technical backgrounds, provided the mentee has a strong interest in this kind of work.
Both mentees should have a strong understanding of the technology, and ideally of agentic AI specifically. Specific coding skills are not required.
Mentees should be able to work independently, identify key uncertainties and make a concrete plan for resolving them, including by gathering external input (with the mentor's guidance).
Time commitment
10h
Application question(s)
- Provide a link to one short writing sample (can be as short as one paragraph), this is to see how you reason, synthesize complex information and communicate. Please do not submit more than a page, and indicate the context behind the work and its intended audience.
- What, according to you, are the key issues that a new fundamental right should be aimed at mitigating? If not through fundamental rights, what other ways would you suggest tackling these issues? I am keen to read your own thoughts. [200-250 words. You can use AI, but please be advised that overrelying on it tends to yield generic answers.]
- Why are you interested in this specific project? [max. 150 words. Feel free to use bullet points.]
- How do your knowledge and skills align with the needs of this project? [max. 150 words. Feel free to use bullet points.]
- What are you hoping to get out of participating in this project (personally or career-wise), and how can I, as a mentor, contribute to it? [max. 150 words. Feel free to use bullet points.]
About the mentor
Elsa is a legal scholar with an interest in agentic AI governance, and a broader focus on designing proactive architectures that ensure AI systems remain safe, legible, and aligned with human economic and legal structures.
In her current role at the Ada Lovelace Institute, she works on UK & EU AI policy. Prior to joining Ada, Elsa was a Summer Fellow at the Center for the Governance of AI (GovAI), where she focused on agentic AI governance. She has participated in and mentored for various policy and technical research programmes in AI, and has previously worked in compliance at a Swiss private bank.
Elsa holds a Bachelor of Laws from the University of Geneva and a Master of Laws from the University of Zurich. She also holds an LLM in International Business Law from King’s College London and completed the Law Foundation Course (conversion to English law) at BPP University.
Note that the projects presented as part of SPAR are entirely independent of her work at the Ada Lovelace Institute.