We are a nonprofit evaluating AI risks in cybersecurity and their policy implications. We seek a colleague to transform our internal research into globally recognized knowledge and drive industry change.

Our outreach strategy is as follows:

1. Draw attention with flashy demos of scary AI capabilities: A kidnapper uses AI to fake a mother’s voice and snatch a child in Texas (target: media).
2. Follow up with rigorous scientific argument: We develop a universal jailbreak that defeats all currently known safety methods at a modest token cost. Pay 1$ in token overhead to make all GPT-4 queries unsafe. (target: technology policy advisors).
3. Ground policy suggestions down into economic analysis: +1% AI improvement on HackBench translates to +$20m losses in the banking sector in 2025 (target: policymakers).

We expect you to build up our research findings into content for each of the steps above. To this end, you will:

• Develop our Twitter media strategy.
• Convert internal research into engaging, memetic content.
• Build strong connections with thematic media (like WIRED).
• Review and edit researchers’ arXiv and conference manuscripts for clarity and concision.

Key skills:

• Strong communication skills: you will need to articulate complex technical and policy-related concepts in crisp, clear language.
• Format versatility: we target academic journals, Reddit, and everything in between.
• Creative thinking: sometimes the best way to make a splash is to run an AR game instead of submitting a conference paper.
• Technical aptitude and curiosity: you need to understand AI and cybersecurity to communicate about them effectively.
• Industry networking: establish connections with leading thinkers.

Compensation: $2000-$3000/month depending on fit. We can go higher for outstanding candidates.

We’re primarily looking for a full-time team member. We will also consider applications for narrow contract work from outstanding candidates. This is a remote position.