[December 2023 edit: This project lives on at alert-team.org.]

[October 2022 edit: See here for our current progress.]

Motivation

See Hinges and crises for expanded explanation.

The hinge of history

So far, long-termist efforts to change the trajectory of the world focus on far-off events. This is on the assumption that we foresee some important problem and influence its outcome by working on the problem for longer. We thus start working on it sooner than others, we lay the groundwork for future research, we raise awareness, and so on. 

Many longtermists propose that we now live at the “hinge of history”, usually understood on the timescale of critical centuries, or critical decades. But ”hinginess” is likely not constant: some short periods will be significantly more eventful than others. It is also possible that these periods will present even more leveraged opportunities for changing the world’s trajectory.

These “maximally hingey” moments might be best influenced by sustained efforts long before them (as described above). But it seems plausible that in many cases, the best realistic chance to influence them is “while they are happening”, via a concentrated effort at that moment. Some reasons for this: there are decreased problems with cluelessness; an increase in the resources spent on the problem; actual decisions being made by powerful actors, and so on.

Crises as times of opportunity

Even if a specific crisis is not a “hinge of history”, crises often bring opportunities to change the established order. For example, policies well outside the Overton window can suddenly become live options, and intellectual development in disciplines and technologies (think of the sudden intense focus on reproducibility in epidemiology). These effects often persist for decades after the initial event (think of taking off your shoes at the airport; think of face masks in post-SARS countries) - and so shaping the response is of interest to longtermists.

How to impact hinges or crises

Acting effectively during hinges or crises may depend on factors such as "do we have a relevant policy proposal in the drawer?", "do we have a team of experts able to advise?" or “do we have a relevant network?” It is possible to prepare these! 

This document proposes the creation of an “emergency response team” as one sensible preparation. 

We tested the above principles during the COVID pandemic, by launching a good-sized research and policy effort, Epidemic Forecasting (EpiFor). We had some success, with associated members advising legislators and international bodies and the associated research getting into top journals and so reaching millions of people. 

Some takeaways:

• There are various paths to impact. COVID illustrated an opportunity for small teams: there are worlds where what's needed is being able to think clearly and do research really fast.
• Often our main bottleneck was project managers, particularly people with both research skills and PM skills. Also EpiFor was only effective because core members already knew each other from FHI RSP or CZEA. And we could have been much more effective, if the team had "trained" together before COVID, to sort out differences in management styles, communication, commitment, etc.
   

How to improve longtermists’ emergency response capabilities

Is it possible to fund, train, and sustain a longtermist response team in the absence of a current emergency? Perhaps - but for people we most want on board, the opportunity costs of being thus “benched” might be too high.

A more viable alternative is a reserve, or standing army: a team of researchers and managers who are “on call” for a future emergency, undergoing annual wargaming or similar refreshers to maintain readiness. 
 

What the team ideally should have:

• existing expertise in many object-level domains
• ability to decompose problems into delegable parts
• structural capital (legal backing, processes, and roles in place beforehand)
• credibility or the ability to signal credibility 
   

ALERT: a rough specification 

(Active Long-termist Emergency Response Team)

What do we need?

• people
• training
• an institution with lots of structural capital
• credibility or the ability to signal credibility 
• a clear trigger
   

People

We propose to gather 30 to 50 “reservists”: people who are able to leap to respond to an emergency. They need to commit to fast activation - something like “given 2 days notice, I am able to switch to this work with >90% probability”. 

Together they should cover the usual object-level domains, and also policy, ops, and narrower domains, e.g. regulators, hardware engineers, clinical trial PIs, lawyers and specialised counsel, NGO managers and directors, media.

We then create some structure - perhaps they form teams of 5-10 people centring on one domain, or one vertical.

Key abilities: <learn anything fast>, generalist research, stats, data science, ML, bio, policy, international relations,...

• domains: bio, medical, ML, policy, international relations, meteorology/disaster, nuclear/chemistry/weapons experts, …
• general abilities: managerial (project, personal/team), communication (soc. networks, policy makers, journalists), presentation/communication, data science (ingress, presentation online, dataset distribution, SW ops), organizing volunteers (for data collection etc.), statistics and modeling, forecasting (directly, creating questions, outreach), finance/accounting, opsec(?)

These requirements are quite severe, but we think there are enough people. Recently more EA projects have involved consulting, modelling, commercial forecasting, founding organizations, etc., and thus more of us have relevant experience. The obvious categories are academics, consultants, and the self-employed. Also, PhD students are a powerful “reserve army of labour”, since (given advisor approval) they have latitude to switch projects at short notice, often without extra funding. Some grantees may also be interested.
 

Institution

We create a "dormant" institution. It is incorporated, legally up-to-date, and able to receive and send money, hire people, etc. This could take the form of a series of NGOs under US, EU, and some other country's jurisdiction. (The teams can be hosted also in some existing institutions.) It has some amount of liquid capital in its account. It has managers on call and a roster of reservists. It has someone actively watching the world, ready to activate the relevant team when a clearly specified risk threshold and relatively low is reached.

Strong leadership is just as important as having experts. In particular, we need the ability to decompose problems into delegable parts, and more generally we need "structural reserves": all the processes and tools to maintain readiness and absorb more people when we need to. This structure includes coordinators for reservists and volunteers, and an oversupply of managers.

Readiness

The institution organises annual readiness exercises, in sprints of 7-14 days. This brings the team together, to work either on a toy problem or wargame, or on some current minor instance of the problem.

The permanent ALERT staff will track what's going on with members, collecting regular state updates. We could coach reservists towards valuable skills (especially more general ones). 

We could keep a live record of the team's collective network: all the contacts with experts, policymakers, funders, regulators, academia, and influential nodes. (The network is vital to supply credibility, without which the teams cannot influence policy.)

Subproject: list of EAs willing to lend their expertise

As well as the reservists, we should have a longer list of people who can help in some way / candidate reservists.

This list would be private (but that it exists would be public or semi-public, as well as the criteria involved).  People who may be willing and able to assist in their domain in a crisis or sufficiently impactful/urgent project

What emergencies?

Where can this team help? Where should they? Where shouldn’t they?
 

• The bar should be high for activating ALERT;
• The main value of the project may well come from presently unexpected events.
  • But they may be correspondingly hard to identify and coordinate on 
    (e.g. does a diplomatic proposal to loosen a nuclear treaty, where there is some chance for intervention via modelling and scientific argumentation with some parties, qualify?)
     

Real-world and hypothetical examples

Good examples

• The Covid-19 pandemic - e.g. early scenario forecasting and publishing information, consultancy for organisations / govts, research and modelling, maintaining datasets (e.g. countermeasures OxGCTR is still very buggy today), developing tools, advising policymakers
• RAMP - an ad-hoc UK academic effort to review hundreds of COVID preprints within days of their release, to aid the civil service in processing the wave of evidence and pseudoevidence.
• Massive geomagnetic storm

Uncertain 

• A rapid vaccine development program (for official/wide deployment) - depends on the timing and possible impact, but the bulk of work in deploying medications seems to be in handling regulatory aspects (incl. research aspects), having narrow-scoped seniority and expertise, manufacturing & supply chain, logistics, deals with target govts etc. - while it may make sense for some EAs to engage, the bulk of work is likely outside of ALERT scope.
   

Negative examples

• Humanitarian crises after local/regional natural disasters 
  (EAs do not have much of a comparative advantage, the skills required are different, and there are many orgs already specialising in this.)

What next?

The first role to fill is the secretary, the person who watches the world, maintains the basic institutional requirements, organises the readiness exercises, etc. Please apply here.

See also

• Holden Karnofsky’s attempt to gather the willing  

This post is intended to be readable as stand-alone, but also forms part of a series explaining my part in the EA response to COVID, my reasons for switching from AI alignment work for a full year, and some new ideas the experience gave me. We co-wrote it with Gavin Leech. Thanks to Nora Amman, Ben Pace, Max Dalton and Tomáš Gavenčiak for helpful comments on the draft.