Project summary

AI-enabled power concentration currently lacks detailed threat models which rigorously map out concrete pathways that lead from today's systems to entrenched control, what capabilities are required for each of these pathways, and how these different pathways and threats interact with each other. A lot of threat modelling is underspecified and narrowly scoped, in terms of the concrete pressures, incentives, and mechanisms, and also in the interplay between domestic consequences and outlooks and the broader strategic and security landscape.

This project is an attempt to fill this gap and create detailed models of how exactly AI could allow a small set of actors to gain a decisive strategic advantage over the rest of the world. I intend to spend 9 months driving an effort in this direction, drawing upon my own knowledge of technical AI safety, institutional design, nuclear verification, and international politics, and my connections with experts in relevant fields, in order to identify the pathways and relevant cruxes, evaluate the likelihood of these pathways and identify the possible outcomes, and map out possible interventions to defend against these pathways.

What are this project's goals? How will you achieve them?

I intend to spend approximately 50% of my time, frontloaded into the initial 4-5 months, on creating and refining detailed threat models with concrete pathways, mechanisms, and quantified assessments of likelihood and impacts. About 30% of my time, mostly in months 4-8, will go into mapping existing and designing potential defenses for the specific pathways. About 20% of my time will go into engaging other people in the field and in adjacent expert communities, in order to get feedback, ensure minimal duplication of effort, and build a shared picture, closing with a small workshop.

At a minimum, I will publish six pieces of public-facing writing analysing specific facets of different pathways to power concentration, plus a comprehensive document detailing the pathways and potential defenses, circulated with relevant people in the community and shared publicly if doing so is not too risky. I will lead the project at half of my time, with a part-time collaborator recruited once the threat-model skeleton exists; for relevant expertise, I can draw on my existing collaborations with the Alva Myrdal Centre and the Oxford Martin AI Governance Initiative.

Theory of impact

AI-enabled power concentration seems to be one of the most critical threats we might face over the coming decade, and one of the ones we are least prepared for. A small set of actors seizing complete and entrenched power would vastly limit the ability of human civilization to prosper and flourish. In addition, the quest to gain this form of decisive advantage would also likely lead to reckless racing, increasing the risks of loss-of-control and potentially great power conflict.

My work is valuable as detailed threat models are the necessary inputs for understanding the nature of the threat, and identifying and prioritizing the potential interventions. Hence, the outputs of this project will be useful to funders and people working in the field, and would contribute to building an actionable agenda in order to rapidly scale future work in an effective manner.

How will this funding be used?

This is the full nine-month programme. My own time is priced at half, at $4,000 per month, because my crisis-instability programme runs in parallel with its own page, and the two pages together fund one full-time person.

The programme runs nine months at $94,000 in total:

• Lead salary: $36,000 ($4,000/month, at half of my time)

• Part-time collaborator (~0.5 FTE, recruited around month 3): ~$28,000

• A small workshop convening relevant people: ~$20,000

• Research compute, tooling, and travel: ~$10,000

Minimum, $14,000: three months of my time at half time plus a compute margin, funding the two central crux pieces.

Who is on your team? What's your track record on similar projects?

I will work solo on this project at the current stage.

• Evaluating and Understanding Scheming Propensity in LLM Agents (arXiv:2603.01608), co-authored at LASR Labs, supervised by David Lindner (Google DeepMind).

• Applying LLMs to Political Challenges in Nuclear Verification (in prep., with Giacomo Cassano, Uppsala). Presented at the Politics of Verification in Frontier AI workshop (Stockholm, April 2026) and the AMC Multidisciplinary Conference 2026.

• Opportunities and Risks from Advanced AI in Nuclear Verification, forthcoming chapter with Sophia Hatz in the Routledge Handbook on Nuclear Verification.

• CVD-for-AI working paper at Pivotal Research with Robert Trager (to be published).

• Ongoing collaboration with the Oxford Martin AI Governance Initiative on strategic parameters for AI agreements.

• Contributor to ControlArena (UK AISI control evaluations) and the HCAST benchmark (arXiv:2503.17354) at METR.

• Public writing on AI and nuclear governance at thenextfrontier.blog.
  

What are the most likely causes and outcomes if this project fails?

The main downside risks are that some of the work might be dual-use in nature and could potentially provide a plausible playbook for concentrating power, and there is a risk of politicization. In order to avoid this, I will extensively red-team outputs, seek private feedback from trusted experts, and redact anything too sensitive to disclose, restricting distribution.

How much money have you raised in the last 12 months, and from where?

Nothing yet. An application for the full programme is pending with Longview's Extreme Power Concentration RFP, and this project is included in the Lightcone Commons funding round that closes on August 23. I will update this page as soon as any commitment lands.