[Cause Exploration Prizes] Creating a “regulatory turbocharger” for EA relevant policies

By Open Philanthropy @ 2022-08-11T10:42 (+5)

This essay was submitted to Open Philanthropy's Cause Exploration Prizes contest and posted to the Forum with the author's permission.

If you're seeing this in summer 2022, we'll be posting many submissions in a short period. If you want to stop seeing them so often, apply a filter for the appropriate tag!

The initiative in a nutshell

Quite many discussions on EA relevant topics forget about the regulatory dimension or show a very low level of regulatory knowledge. The designed initiative “regulatory turbocharger” could enhance many EA relevant policies by regulatory knowledge collected over decades in many regulatory fields and by strengthening the regulatory dimension of the policies. Regulatory tools and measures could thereby strengthen EA relevant policies. The regulatory knowledge is thus used as a means to increase the effectiveness of EA relevant policies, in the same way as a turbocharger increases the power of an engine without requiring noteworthy additional fuel.

Why is the regulatory dimension relevant for EA topics?

a) Many directly life-saving policies like anti-violence, anti-tobacco or anti-alcohol policies need underpinning by regulation.

This is evident for policies regarding tobacco, alcohol and other drugs: just by campaigns many actors will not change their business activities. The many aspects that can only be covered by regulation are listed in these three articles (1, 2, 3) and this model law. 

The Regulatory Institute has published various articles that exemplify how regulatory tools can and should be used to fight violence. See example given this article regarding sexual violence or the articles regarding the protection of minors and of older persons. See as further example the fight against the use of fire-arms for which the Regulatory Institute has published this article.

b) Regulation is key in many policies that have a life-saving side-effect and that reduce human suffering, like anti-noise policies, anti-environmental pollution policies, traffic safety policies, medicines and medical devices authorisation, and technology authorisation

Noise is causing fatal heart diseases. It is mainly addressed by regulation. Regulation limits the maximum noise emission of engines (e.g. in cars) and sets up behavioural rules, namely at certain hotspots, see the speed limitations established to reduce noise.

Modern environmental policies use both programs or campaigns aiming at behavioural change and laws / regulation to pursue their goals. In particular the not so well-minded populations will only react to laws / regulation.

The same applies to traffic safety policies. The fabulous reduction of mortality rates in France over the last three decades can only be explained by tougher rules and their strict enforcement, which even led to a complete change of driving style of the French population: whilst, 30 years ago, the French used to drive very dynamically and thus running high risks, their driving style has become similarly calm as the driving style of British or even American drivers.

Regulation is also key for the admission of medicines and medical devices to the market and for the state reimbursement rules with regard to them. Complex risk-benefit considerations and risk minimisation principles need to be laid down in law to trigger a harmonised practice, to reduce incidents and to maximise the medical benefit. But even for simpler products or technologies where there are only risks and no noteworthy counterweighting benefits, regulation is needed to minimise risks.

These are just some of the many policy fields with live-saving side-effects that can be strengthened by regulation. A more comprehensive analysis would detect several dozen more.

c) Protection of animals requires clear and enforceable regulation

Regulation offers manifold possibilities to protect animals and even to increase animal welfare, see these two articles (1,  2) or even to create conditions where the suffering and killing of animals is not needed anymore in the first place, for instance via the development of cultured meat. Not using the possibilities of regulation implicitly weakens animal protection policies.

d) Artificial intelligence policies are often regulatory policies

More and more rich jurisdictions are developing or considering regulation on artificial intelligence, whilst the existential risk dimension is often still to be considered to be a taboo (N.B.: existential risk is here defined as risk for the entire mankind). Possibilities for regulating artificial intelligence are covered by these two articles (1, 2) and this model law. For most of the issues to be tackled in the field of artificial intelligence, there are no equally efficient alternatives to regulation: codes of conduct or other forms of self-regulation do not trigger the same level of commitment and compliance.

e) Existential risks can partly be countered by national regulation and international agreements

The existential risk dimension of artificial intelligence is not the only example for which we can state that a purely national approach is unlikely to be fully successful. Other examples are certain technology and research risks, like the spreading of uncontrollable micro-organisms developed in laboratories. The article “The Need for Global Protections Against Existential Risks” provides further examples.

But to cover efficiently existential risks, both the international and the national playing fields need to be served. International obligations without transposition at national level by regulation / laws will not suffice. Both need to go hand-in-hand.

How could a regulatory turbocharger operate?

The regulatory turbocharger could target private and public organisations developing policies for EA relevant topics, and subject to available resources even governments and parliaments, by the following activities:

a) Suggesting underpinning regulation

Quite some private or public organisations, and even often governments and parliaments, are unaware of the regulatory dimension of their topics or shy away from that dimension as it seems too cumbersome or technically difficult. The regulatory turbocharger could explain the added value of regulation. The regulatory turbocharger could also reduce the fear of approaching the regulatory dimension, by showing-up simple regulatory tools. Finally, the regulatory turbocharger could reduce one of the main obstacles for the use of the regulatory dimension: the workload linked to the adoption of regulation. In particular, the regulatory turbocharger could point to existing model laws. Model laws will soon be made publicly available in a comprehensive, structured model law library set up by the Regulatory Institute. In addition, the regulatory turbocharger could help to adapt the model laws to the precise situation that the organisation wishes to tackle.  

b) Suggesting effective control and sanction mechanisms

Even where the regulatory dimension as such has been acknowledged so that regulation is on the agenda of governments and parliaments, there is plenty of margin for improvement. According to the experience of the Regulatory Institute, less than 1% of laws / regulations are complete in terms of effective control and sanction mechanisms; see as one of the few examples the Ugandan Anti Money Laundering Act 2013 and compare it with the mostly very much incomplete other laws mentioned in this article. In that context, it is important to recognise that the knowledge regarding effective control and sanction mechanisms is not linked to a specific sector, whilst some sectors have been pioneering in inventing particularly effective control and sanction mechanisms. Based on the knowledge collected by the Regulatory Institute, e.g. in this checklist, the regulatory turbocharger could inject best available regulatory techniques from any sector into the EA relevant policies. 

c) Identifying incentives that can be implemented by regulation, in addition to classic obligation and sanction schemes

Where organisations, governments and parliaments are aware of a regulatory dimension of their policy, they often fall into the opposite trap: seeing everything in terms of classic law, classic law being based on obligations and sanctions. However, there are many incentives that can be used in regulation to reach a certain policy goal which go beyond obligations and sanctions. Reputational advantages and disadvantages, faster and easier procedures, and exemption from the obligation to renew a licence are just some of the many incentives that can be used in regulation to reach a certain policy goal. The regulatory turbocharger could promote EA related policies by highlighting these incentives, namely where they can be implemented by regulation. The policy goals can be better reached by these additional incentives.

d) Suggesting alternative supervision and enforcement mechanisms, even involving private natural and legal persons

With the increasing technological and societal complexity of our world, there are ever more policy fields which need to be covered by regulation, whilst the manpower of administrations called upon to implement these policies is rather shrinking, due to restrictive budgetary policies. Hence the public authorities’ enforcement of regulation in general, but also for policies with relevance for EA, is getting weaker from year to year. One approach to counter this mismatch is to involve third parties and even private persons in the enforcement of laws / regulation; see as an example this proposal for a regulation which obliges importers and distributors to verify certain basic elements of compliance of products for which the respective manufacturers are responsible; see as another example the practice of the city of New York to offer money to those private persons documenting by videos the running of engines of parked trucks (a problem for the environment and for human health both in terms of emissions and noise) and testifying on the running engines at courts where the indicted lorry drivers launch legal appeals. 

e) Developing model provisions and model laws

Even though the Regulatory Institute has already developed a few model laws, including one on the EA relevant topic of artificial intelligence, one on emergencies and one on research and technology risks which all three might cover certain existential risks, there are many EA relevant policies which are not yet covered by good quality model laws or even just model provisions that can be embedded into national regulation. Given that many organisations and even the governments and parliaments are overburdened with the task of developing laws / regulation, model provisions and entire model laws are a key instrument for levering the regulatory potential in EA relevant policy fields. The fact that so many jurisdictions cannot cope with all regulatory challenges cumulatively is also visible in the fact that for most policy fields investigated so far by the Regulatory Institute, less than 50 nation states have adopted any kind of law / regulation. On a - still to be developed - matrix listing all the nation states and all the policy fields for which laws / regulation are useful, the vast majority of fields (we estimate: > 90%) are blank. EA relevant policies are no exception to this rule. Hence model provisions and model laws can lower the hurdle for jurisdictions to use law / regulation as a policy means, and the same for the private organisations which have an interest in the development of policies in and for the respective jurisdictions.

f) Preparing standard provisions that can be used in international (bilateral or multilateral) treaties to promote effective altruism topics

The article “The Need for Global Protections Against Existential Risks”, published in the “Regulatory Review” of the Penn State University, exemplifies how standard provisions can be used even to tackle the most difficult international policy challenges, where sacred principles of national sovereignty, national pride other national interests are at stake. Another example is the fight against terrorism and money laundering. The respective international convention which contained model provisions in its annex has led to a wave of enhancement of national laws across the globe. See for details this article. 

But both articles only tip on the top of the iceberg in terms of what could be developed over time by a regulatory turbocharger: regulatory mechanisms to overcome the “tragedy of the commons” in international politics, the tragedy of the commons being the root cause for decades of failed efforts to stop climate change and other unsolved international problems, including those triggering existential risks. The top potential to be levied by the regulatory turbocharger might be to develop regulatory mechanisms, vinculating certain advantages of international cooperation, e.g. free trade or financial assistance, to a strict existential risk minimisation policy which is also transposed by domestic regulation / laws. 

g) Adapting on request regulatory tools to the concrete situation in which the policies are applied in a certain jurisdiction

Finally, there is always an important potential for improvement by adapting regulatory tools to the concrete situations in which policies are applied in a certain jurisdiction. The adaptation to that concrete situation requests some investigation on the ground. The socio-economic situation, the mentality of the persons concerned, their education and societal constraints need to be taken into account to optimise the effect of legal provisions. Having a focus on EA relevant topics, plus the cross-sector knowledge on regulatory tools and, hopefully, the human resources necessary to investigate the concrete situation of jurisdictions, the regulatory turbocharger would be in a position to really optimise laws / regulation. The importance of that optimisation can in particular not be overestimated when it comes to existential risks.  

How could a regulatory turbocharger be set up organisationally?

There are many organisational possibilities. In the following, five possibilities are presented in the order of size of the respective undertaking:

a) Sending a Regulatory Institute staff temporarily into an EA organisation, that EA organisation hosting the regulatory turbocharger

This possibility is useful where there is only a very limited budget available, but where staff of an EA organisation can also be used. Currently, the Regulatory Institute could make available at best a fourth of a full-time equivalence where there is no financial compensation and half a full-time equivalence where there is financial compensation. This possibility is only meaningful where there is ensured personal continuity on the side of the EA organisation as otherwise the regulatory knowledge transferred will be quickly lost again.

b) Regulatory Institute to receive staff from an EA organisation for training so that this staff can inject regulatory knowledge later-on from its position in the EA organisation

This possibility requests some investment from an EA organisation: staff working for the EA organisation would be made available to the Regulatory Institute for a certain period of time in order to get trained on regulatory techniques.  This possibility, too, is only meaningful where there is ensured personal continuity on the side of the EA organisation as otherwise the regulatory knowledge will be quickly lost again.

c) Distinct team within the Regulatory Institute

Establishing a distinct team within the Regulatory Institute would require additional financial resources. Even with additional financial resources, this possibility would currently be limited to two persons working part-time, as otherwise the internal organisational equilibrium of the Regulatory Institute would be too much disturbed.

d) Distinct team within an EA organisation, established with support of the Regulatory Institute

A slightly higher investment is needed where a distinct team is established within an EA organisation. The Regulatory Institute could contribute by training and supervision of activities of the regulatory turbocharger. Supervision would consist of commenting on draft position papers and strategy documents. 

e) Creating a distinct legal entity that can, contrary to the Regulatory Institute, deal exclusively with topics of high priority for the EA movement

An even higher investment is needed to make this possibility meaningful. At least two full-time equivalences are needed to set up a distinct organisation. The distinct organisation could be established as a daughter entity of an EA organisation or as a joint venture of the Regulatory Institute and an EA organisation.

What would be the cost-benefit ratio in case of a medium-size organisational set-up?

Much depends of course on the resources input and the respective organisational set up. We describe in the following a medium-size organisational set-up and equivalent output. 

•   Cost example: 2 staff working 50 weeks at 10h/week at (on average) 30$/€ per hour = 15.000$/€ / year

•   Corresponding output example:

• Injecting >20 times regulatory aspects/tools into discussions in EA fora
• Contributing to 3 position papers of animal protection organisations by inserting regulatory control mechanisms;
• Suggesting enhancements of the anti-tobacco legislation of a 40 million inhabitants state with 8 million smokers;
• Disseminating standard provisions for national laws + international agreements compelling all to counter existential risks. 

•   Cost-benefit ratio based on the last output alone:

• Likelihood of this initiative hindering human extinction: 1/1.000.000;
• Potential benefit if extinction can be hindered: 75 years human life x 100.000.000.000.000 future humans (most recent “Kurzgesagt” estimation disseminated in latest EA newsletter) = 7.500.000.000.000.000 human life years;
• Potential benefit multiplied by 1/1.000.000 likelihood of success = 75.000.000.000 life years;
• Ratio: 75.000.000.000 life years for 15.000$/€; equivalent to 5.000.000 life years per 1$/€.

 •   Cost-benefit ratio based on the last but one output alone:

• Likelihood of this initiative influencing the legislation of the 40 million inhabitants state with 8 million smokers: 1/10;
• Potential benefit if the anti-tobacco legislation is enhanced: 1 out of 100 persons that otherwise would become a smoker does not become a smoker; accordingly, there will be 80.000 smokers less over 20 years whilst smokers have according to various scientific views between 7 and 14 years shorter lives; we take 10 years as a basis; hence 800.000 human life years could be saved;
• Potential benefit (800.000 human life years) multiplied by 1/10 likelihood of success = 80.000 life years;
• Ratio: 80.000 life years for 15.000$/€; equivalent to 5.3 life years per 1$/€.

How could this initiative be scaled up?

Once the regulatory turbocharger has been successfully been set-up, staff of EA relevant organisations could be invited for traineeships so as to learn how regulation can be used to enhance policies in EA relevant topics. The knowledge would thus flow into the organisations developing policies in EA relevant topics. Moreover, the regulatory turbocharger could offer training for staff of EA relevant organisations.

Is there already anything alike?

There is nothing like a regulatory turbocharger for EA relevant policies except to some extent the Regulatory Institute. The Regulatory Institute has worked with certain governments and parliamentarians with regard to EA relevant policies, not least in the fields of anti-tobacco, anti-alcohol, environmental, and artificial intelligence policies. However, the Regulatory Institute cannot and will not with its scarce resources engage with NGOs and other organisations specialised in EA relevant policies - it even does not fully cover its current audience, governments and parliamentarians. In addition, it needs to re-equilibrate its focus topics. So far, EA relevant policies happen to have been covered disproportionately often by articles, model laws and corresponding reach-out activities. In view of maximising public utility and improving the reputation of the Regulatory Institute as neutral, non-activist knowledge platform, it is necessary to follow more closely the demand of governments and parliamentarians and thus to turn to other topics, namely emerging new topics like the metaverse for which exerting some influence is easier than for well-known, old topics. This means unfortunately that  less of the Regulatory Institute’s very limited resources will be made available for EA relevant policies. The gap is thus likely to widen over time, not to shrink, unless something like the regulatory turbocharger is established.

The absence of regulatory advice for organisations active in EA relevant policies would not harm if governments and parliamentarians had by themselves the regulatory knowledge necessary to develop regulation which is complete in terms of regulatory tools. Astonishingly, very few jurisdictions collect regulatory tools knowledge in a way that merits the term “systematic” (we know only of limited knowledge collection by Australia, Switzerland and  France). Mostly, jurisdictions only collect and disseminate knowledge with regard to drafting modalities and styles, which is a different issue, but unfortunately an issue that conceals the need for knowledge on regulatory tools. Some jurisdictions and international networks refer to the Regulatory Institute as regulatory tools knowledge storage, but this reference alone does not guarantee that their respective laws / regulation is complete in terms of regulatory tools. As already stated, less than 1% of all laws / regulations are complete when measured against what is possible and useful. Accordingly, the absence of regulatory advice for organisations active in EA relevant policies is not compensated by inherent knowledge of governments and parliamentarians.

The author of this concept and potential contributor to its implementation: Manfred Kohler

•   Studied law in Germany and public management in France (ENA);

•   Working for national and international administrations for three decades, in law-making for two decades; working for or with NGOs for three decades;

•   Co-drafted > 200 regulations for UNECE and the European Union;

•   Gave trainings on regulatory techniques at the European Commission;

•   Founder of various organisations, including of the Regulatory Institute;

•   Author inter alia of the Regulatory Institute Handbook “How to regulate?”, of Research and Technology Risks: Part IV – A Prototype Regulation, Model law on emergency management, and of an article in the “Regulatory Review” of the Penn State University: “The Need for Global Protections Against Existential Risks”;

•   Would be available as from March 2023.