One thing I'd add to this is that there can be a culture of policy entrepreneurship. The twitter neoliberals in particular are really good at shipping round policy ideas, improving them as they go. EAs are much worse at this, for reasons I don't understand.
Policy entrepreneurs are "individuals who exploit opportunities to influence policy outcomes so as to promote their own goals, without having the resources necessary to achieve this alone. They try to create new horizons of [policy] opportunity through innovative ideas and strategies.[1]"
It is a process that involves individuals who are willing to take risks, identify policy problems and solutions, and use their political skills and timing to achieve a specified outcome.[1]
This post is a valiant attempt to convince someone that policy entrepreneurship is a vital investment of their 80,000 hours. Should they wish to use them to advance EA-adjacent causes like AI safety, ending factory farming, or pandemic prevention.
I'll explore ideas about how we can shape the future from Will MacAskill's WWOTF - including plasticity and the importance of early intervention.
I'll also introduce the inner/outer policy landscape dichotomy. I argue that EA doesn't invest enough in shaping the outer policy landscape. I'll explain the concept of the party political retail offer, and how a policy entrepreneur might go about shaping it.
This article has a somewhat patchwork writer's context. Partially comprised of scribbles in the margins of my notes as I work through the EA Cambridge AI Governance reading list. Partially comprised of sticky note scribbles from my ever more annotated copy of WWOTF. Both of these contexts shape my approach to making this case, so helpful to include in this section.
Consigned to my iPhone Notes app, until I read Will MacAskill’s reflection on Bill McKibben's recollections of the history of climate change activism on page 43 of WWOTF:
Bill McKibben: "We haven't taken [advanced artificial intelligence][2] seriously because it doesn't, at the moment, impinge on our day-to-day life. But one of the things climate change taught me is that things happen fast, like, really fast. And before you know it they're out of control. So the time for thinking about them is when there is some chance of getting a handle on them."
Will MacAskill: "... with climate change, we may have missed one moment of plasticity, and we should hope there are more to come. But perhaps we can also learn a more general lesson and respond more rapidly to new challenges - list of examples - as soon as they arise."
EA doesn't have a strong affinity for politics in the electoral sense. Where it does attempt to articulate a theory of policy change, it tends to focus on working politically within the big institutions - the EU, UN, WEF, NATO, the DC machine etc- or as a civil servant.
This conception of policy entrepreneurship relies on some people doing that. It is vital work. For the sake of shorthand, I'll refer to this as the inner policy landscape. As in, inside institutions.
Where I go further than EA, in general, is in my assertion that this is a chair with two legs. This kind of policy work isn't sufficient in itself for the chair to stand.
In order for people within institutions to advance good[3] policy beyond white papers and conceptual studies on the kinds of issues EA concerns itself with, with the urgency they demand, those issues need to advance up the policy priority tree.
To do that, you want to shape the retail offer - which is shaped by the outer policy landscape.
Policy priorities, and the urgency of policy implementation, are shaped by the political environment the government is operating in.
What's on the news? What are people rowing about at the pub - or protesting about in the street? What are the swing voters thinking about? What was in the manifesto that got them over the line vs their opponent at the ballot box?
What issues are shaping the great public debate - and what can I put on the leaflets and the telly[4] that will move the masses in my direction at the next election?
For the sake of shorthand, I'll refer to this as the outer policy landscape.
Key point: EA invests a lot of time and resources into shaping the inner policy landscape - but very little into the outer policy landscape.
There's a somewhat detached view within the EA community that shaping policy happens within expert advisory. It's an aspirational idea if nothing else. This isn't an argument against people investing their careers in becoming experts and trying to influence policy that way.
But, I’d argue[5], that the retail offer has an outsized impact that we’re undervaluing in our current account of policymaking. Which, as a knock-on, means we're undervaluing this as a way to spend your careers doing good in this field.
After all, we aren’t short of climate experts and expert policy proposals - but we’re chronically short of impactful passed legislation.
The ‘retail offer’ of any given political party has an outsized impact on the priorities of governments and the scrutiny they face. Be that tuition fees in the 2010 Liberal Democrat manifesto, or Jeremy Corbyn’s plans to nationalise utilities and everything else while he was at it in 2019.
Not everything in a manifesto is the retail offer. It’s the handful of simply expressed core policy claims a party makes in the media and election messaging in its bid for power.
The retail offer is shaped, in order of importance, by the court of public opinion + the news cycle + ideology + whatever niche issue the leader and his inner circle have their eye on + how much cash they’re expecting to find in the proverbial attic.
The court of public opinion and the news cycle - which both have an outsized impact - are fields that policy entrepreneurship has the greatest possibility to influence. They reside firmly in the outer policy landscape.
The rest is broadly ideology. How soon you lock down in a global pandemic depends more on how your ideology encourages you to weigh personal liberty against community welfare than anything your scientific advisors tell you. It’s as much an existential gamble (will my core vote vote me back in again if I do this) as it is an existential gamble in the traditional sense.
What the government promised its voters shapes what the government does. With whatever resources it has left over while handling whatever urgent nightmare scenario it is faced with in our interesting times. If you want preemptive planning for a potential crisis - rather than reactive firefighting - it needs to make the retail offer.
Returning to the WWOTF quotes that inspired this article in the summary:
Bill McKibben: "We haven't taken [advanced artificial intelligence][2] seriously because it doesn't, at the moment, impinge on our day-to-day life. But one of the things climate change taught me is that things happen fast, like, really fast. And before you know it they're out of control. So the time for thinking about them is when there is some chance of getting a handle on them."
Will MacAskill: "... with climate change, we may have missed one moment of plasticity, and we should hope there are more to come. But perhaps we can also learn a more general lesson and respond more rapidly to new challenges - list of examples - as soon as they arise."
Things happen fast. If we accept that, then the payoff for being ahead of the curve to the best of our ability is significant. The speculation in the same chapter as these quotes on what the trajectory of the world in regards to climate impact would be had we instituted a carbon tax earlier is a strong case for believing this line of argument.
There’s a particularly important offhand reference to political polarisation in this section of the book worth picking up on.
It would, I'm confident in my surety, be easier to pass climate legislation based on good climate science and the urgency of the crisis at a time where climate change hadn’t become ‘culture war’ value clustering shorthand. There it currently resides, alongside issues like gender, race and sexuality liberation, as a shorthand for leftism and the liberals.
Polarisation being a force for less than ideal policy is an idea the community already considers worth tackling/investing in.
Building cross-factional consensus, and a political climate that is conducive to delivering your policy perspective of choice, is a big part of policy entrepreneurship we currently neglect. It's outer landscape stuff.
Will MacAskill introduces the concept of plasticity, in relation to our capacity to shape the future, with this quote:
"Often, some event can have highly significant, persistent, and contingent effects if there is a period of plasticity, where ideas or events or institutions can take one of many forms, followed by a period of rigidity or ossification."
"The dynamic is like that of glassblowing: In one period, the glass is still molten and malleable; it can be blown into one of many shapes. After it cools, it becomes rigid, and further change is impossible without remelting."
He goes on to give examples of scenarios where he thinks plasticity occurs most often: After a crisis, when some idea or institution is still new, and when setting the laws and norms surrounding new technologies.
MacAskill's plasticity analogy, if you accept it, really hammers home the importance of both early influence and preparing to have influence at the most impactful times.
Moments of plasticity could and probably will be so much shorter than we would like. As Bill says above, you need to start now. At the time of policymaking urgency, you want to be operating in an environment which isn't hostile to a particularly difficult but necessary change (think car use in the climate debate). You want to have won the battles already, to make it easier to win the war.
Nobody wants to be laying the foundations for a much-needed house to shelter from the rain while it’s flooding and they had months of summer during which to start.
With the gift of hindsight comes the gift of time. Time to build grassroots support. Time to build good relationships. Time to build a strong press narrative. Time to position influential political actors in the right direction of travel.
This is work we do not invest in enough.
Borrowing from the ODI think tank's Tools for Policy Impact guide, there are four broad styles of policy entrepreneur: story-tellers, networkers, engineers and fixers. ODI explains it better than I could, so consider this useful background reading.
You'll need all of these kinds of policy entrepreneurs to succeed. Policy making is a team - and very much contact - sport. We don't invest anywhere near enough in placing people in these roles.
Within member-led policy parties, there are real advantages to be had in investing in convincing the grassroots and winning conference votes. Major housing developers don't pay £00000 sums to attend party conferences to make friends. Far cheaper routes include hosting discussion panels, forming internal groups and registered internal organisations and getting enough signatures together to get a policy voted on at a conference.
There are a lot of niche policies that take this route and fail to make the retail offer. But, combined with a favourable media and public interest landscape, you can shape the value approach behind the retail offer policy this way.
The court of public opinion is a great source of urgency if you can convince it. Elected politicians need to be seen to be talking about the issues people care about - from the right perspectives for their core vote.
People care about salient things. Issues that impact them day to day. What their friends talk about at the pub. What they see on Facebook.
Your work, typically at the grassroots level, should focus on making an issue as relevant to as many people as possible and getting it in front of them again, and again, and again... There's an art to this - that I think lots of campaigning organisations miss by a mile. You want to be positioning yourself as a voice of the normal people, not a weird fringe group. This is a trap AI safety and animal welfare campaigners often get caught by.
You're winning when the issue becomes so important to a person that it becomes a vote winner or loser. You can artificially manufacture this sense of anxiety (and urgency) in elected politicians by organising noisy protests and getting on daytime TV and in the niche political press.
The attitude of the press plays a significant role in both of these approaches.
The press is a vast landscape I couldn't begin to explain in the detail it deserves here without it being a whole other post. But, for some broad influence principles:
Politicians, casually, and normal people aren't reading the same outlets. Politicians are more likely to have niche interests or committees/roles where they need to keep up with niche sector news. They're more likely to read the heavy-duty Economist and Telegraph-style stuff than The Sun at home.
Political parties and whips offices are dropping serious cash on the morning papers. They'll try and stratify their collection, across class-associated and geographic lines. They'll have a copy of The Sun next to a copy of the New European. They will make a half-attempt at judging which groups care about which things based on which issues make what papers. A broader reach = a broader perception of support.
It is far easier to get your ideas in front of normal people if you can turn a complex policy idea (like AI safety) into a consumable slice-of-life story. Unlike most of the content on the EA forum, you're aiming for a reading level of 11 years old.
The 80,000 Hours page on party political careers makes some good points in this regard. The likelihood of any given person reaching an influential position where they can enact policy is pretty unlikely. The greasy pole is a hard climb.
It does, however, rate this sort of work as high in advocacy potential.
Returning to the two-legged chair analogy: If nobody is doing this work, the people working in the inner policy landscape are going to have a far harder time trying to deliver good policy. And vice versa. The chair stands when we have a good spread of people doing both.
As with all 80,000 Hours career advice, the name of the game is skill set. If this kind of work is something you would be really good at, over and above a more 'impactful' line of work, this is an avenue worth exploring. From my brief sojourns thus far in the EA community online, this appears to be a skill set the community is chronically short of.
If you're considering a policy-adjacent career, it is worth trying to work out which side of this coin you might be strongest on.
Cohen, Nissim. Policy Entrepreneurship at the Street Level: Understanding the Effect of the Individual. Cambridge: Cambridge University Press. pp. 3–7. ISBN 9781108875233 - An essential read for aspiring policy entrepreneurs.
Added by Will MacAskill in WWOTF for clarity
Effective, value-aligned, broadly positive.
I'm an experienced digital comms strategist and even I cannot convince the powers that be that what will move the needle online should matter in their creation of the retail offer. They're thinking leaflets, telly, and doorsteps.
Drawing from experience on the inside of a manifesto development process (national and local), experience as an elected politician, and from all the best stories about how government really works from former government ministers and current MPs over pints.
From the perspective of a party political insider. I have no doubt that civil servants have a different conception of how this works. The answer is somewhere in the middle.
For those who wish to develop a better sense of politics, you can't go far wrong with following the local political news and trying to get an opportunity to shadow a politician you respect.
You want to get insight into both the meetings and campaigning and the prep work that goes into them.
Many evenings during my undergraduate degree were spent on incredibly uncomfortable public gallery benches at various council meetings - following the proceedings and mapping the politics. This was partnered with assisting an opposition leader with comms and policy projects. Skills, and back pain, for a lifetime.
Some half baked thoughts: I think this post is good and I agree these two models of engaging with policy should be pursued simultaneously. I just think that this post underrates how much of this is already happening, but in a slightly less public and coordinated way, since there is reputation risks for the EA movement if it starts to be associated with political parties / specific politicians.