FAR AI is seeking expressions of interest from freelance science communicators with a range of skills from technical writing, graphic design, presentation development and web/social media management. These science communicators will help effectively communicate the results of research to the AI research community and key stakeholders.
FAR AI is a research organization that accelerates neglected but high-potential AI safety research agendas. We support projects that are either too large to be led by academia or overlooked by the commercial sector as they are unprofitable.
FAR AI aims to reduce existential risk from transformative AI systems. Broadly speaking, our research is focused on how AI can learn about humans’ goals and then robustly help humans achieve them. Our expertise is in reinforcement learning (RL), one of the most general frameworks for building transformative AI. In our prior work, we have developed ways to measure the goals of RL agents and teach a single agent to achieve many different goals. We have also identified a new threat model for attacking RL systems, vulnerabilities in narrowly superhuman systems and developed a method to make RL systems more robust.
We believe that the communication of results is critical to the success of research. Research has shown that language editing improves perceived research quality and even that tweeted papers are better cited. Currently, we have various research projects currently bottlenecked by the communication of results, and are hiring science communicators to collaborate with us.
“At least 50% of the time [on a research project] must go for the presentation. It’s a big, big number.” - Richard Hamming, You and Your Research
About the role
You will communicate results to the research community and other key stakeholders. You will work on a variety of projects led by different researchers. We want this role to help you evaluate your personal fit for, and potentially launch, your career in this field. Your contributions will be recognized with either co-authorship or acknowledgements as appropriate, enabling you to develop a portfolio. We would also be excited for you to pursue independent side-projects, such as starting a blog which could be part of the Effective Ideas ecosystem.
Our current projects need to be communicated in various different ways, which require the skills outlined below. We expect even strong candidates will have experience in only a subset of these areas, and are open to hiring multiple people with different skill sets. So if you think you might be a good fit for some of these areas, please apply!
- Editing drafts of academic papers to more effectively convey the ideas to a scholarly audience.
- Writing blog post summaries of papers, targeted at a technically knowledgeable (but not necessarily expert) audience (example).
- Drafting blog posts from notes of conversations or debates (example).
- Creating illustrations to effectively communicate complex concepts in papers, blog posts or presentations. For example, see the illustrations in the Eliciting Latent Knowledge agenda, or Figure 1 in the Inverse Reward Design paper.
- Creating logos for research projects or software (example logo and its associated codebase).
Presentations and recordings:
- Developing engaging presentations, both for a scholarly (e.g. conference talk) and lay audience (e.g. public presentations). For example, Adam Gleave’s conference talk or Stuart Russell’s Reith Lectures.
- Recording and editing audio and video content, such as research summaries or interviews between researchers.
Websites & social media:
- Creating and maintaining websites, to showcase a research project or lab.
- Helping manage the social media presence of prominent researchers or labs.
You might be a good fit if you’re:
- Enthusiastic about communicating complex ideas as simply as possible.
- Are results-oriented and motivated by impactful research.
- Proactive and can independently prioritise your tasks.
- Are familiar with some of the main themes of technical AI safety research.
Prior experience is helpful but not required.
- Applications: Applications will be evaluated on a rolling basis. The application process consists of a paid work test (8 hours) in which you will write a short blog post summarising a technical AI safety paper, a short non-technical interview, and a paid work trial (~2 weeks). If you are not available for a work trial we may be able to find alternative ways of testing your fit.
- Compensation: This is a contractor role compensated at $30-$50/hour depending on experience. We will also pay for work-related travel and equipment expenses. We offer catered lunch and dinner at our offices in Berkeley.
- Duration: This role is temporary but may lead to a long-term role either with us or the labs you will be collaborating with. This is dependent both on your fit for the role and the demand for communication services from our collaborators.
- Hours: between 10-40 hours/week, depending on what tasks you can perform and how many individuals we hire.
- Location: The position is remote, although we can pay for periodic travel to visit collaborators.
- Org: You will be contracting with FAR AI, and will report to a staff member at FAR AI weekly. You may also work with researchers that we are collaborating with.