To make good decisions on complex issues (e.g. public health), participants need to feel safe to express divisive perspectives and have the listening skills and willingness to consider opposing opinions. During the COVID-19 crisis, the WHO implemented wholesale censorship across both traditional media and social media. Even within the scientific community, open discussion is censored. This top-down control is reducing the variety of discussion and proposals that could potentially be considered. In a healthy ecosystem, multiple perspectives could be considered and tested. The structure of a DAO has potential for better sense-making and richer discussion.

While many social media platforms have caused increased antisocial behaviour, well-designed systems can cause better sense-making. One of the earliest and most long-standing threaded chat platforms, Slashdot.org, included mechanisms for people to indicate the quality of others’ responses to discussions and to acquire reputation over time. Loomio offers a discussion platform with mechanisms that encourage collaboration and safety. More work needs to be done to develop platforms and mechanisms for inclusion that are not driven by market incentives, but rather designed to provide psychologically safe places for thoughtful discussion and deep consideration of alternative viewpoints and ideas. Recently, the emergence of channels such as Rebel Wisdom and The Stoa have shown the public’s desire for in-depth discussion, but these are generally moderated discussions between experts and not designed for the general public to engage in such discourse.

The focus on ‘signalling’ and ‘preferences’ ignores facts and expertise. Intelligent decisions include both facts and perspectives. Factual information must be presented as factual, along with information about the clarity or reliability of the information. Scientific studies and known use cases are different from people’s opinions and perspectives. Perspectives are equally important, however. It may be factual that an infectious disease is fatal, and it may be factual that social distancing is causing a rise in suicide and addiction and having a long-term impact on mental health. Facts and statistics can be presented to decision-makers about all of these impacts, but facts are not sufficient: people’s values determine what result is ‘best’ for them. Different cultures and segments of the population have different values about the importance of these impacts. Decision-makers require both reliable facts and multiple perspectives.

Contemporary research of Dr Anna De Liddo of the Knowledge Management Institute has led to a number of demonstrations of collaboration platforms that help people form better opinions and improve critical thinking. By developing a platform where people must discuss evidence for their claims, her team is looking at how to create a safe environment that allows recognition of expertise and encourages people to understand the content of a claim as well as its source. The Consider.it platform developed by Dr Travis Kriplean offers a discussion platform designed to help people reach a deeper understanding of each others’ viewpoints and provide visualisation to describe the reasoning behind those opinions.

The problems we face as humanity affect different populations in different ways. Depending on your perspective, damming a river could have positive or negative effects. Almost every interesting problem has paradoxes. Problem definition needs to take into account multiple perspectives, and problem definition must be a prerequisite to proposal-making.

None of the DAO platforms to date have capabilities for problem definition. Yet without problem definition, how can a community determine if a proposal has merit?

Communities need a way to define and prioritise the issues to address. Some platforms, such as Canonizer, identify issues based on the volume of discussion and provide intelligence about how divisive the issues are to a community. However, just because an issue is interesting and divisive doesn’t make it a priority. People may feel very strongly about the gender denomination of bathrooms, but most would agree that it is not as important as the curriculum of the school in which the bathroom is located.

If a ballot has only bad or mediocre options, democracy is meaningless. Organisations use multiple methodologies to brainstorm, compose and revise propositions. DAOs today allow anyone to propose anything, but they don’t recognise or reward collaboration or creativity. While platforms such as Aragon and DAOstack encourage a period of informal discussion and deliberation on proposals, it’s not required.

Aragon enables periodic voting schedules, so discussion is conducted over a period of time, and then voting is on a tranche of proposals together. The DAOstack paradigm allows ongoing proposal-making, so people are voting on proposals as they appear, without comparison to past (or future) proposals. This type of yes/no, ‘first come, first served’ proposal-making favours speed and competition over collaboration, deep thought or consideration of minority perspectives. Making decisions this way is like walking down a street and deciding whether to eat at a restaurant without knowing what restaurants are around the corner. You must make a yes/no decision for one option at a time, and if a majority always wins, the person who is vegan may go hungry.

The Holographic Consensus mechanism on DAOstack prioritises popular proposals, but more testing is needed to see if it’s effective. The most popular proposal isn’t always the wisest one.

Distributed technologies have the promise to create a wide variety of solutions for inclusion, but so far, none of the systems in place have demonstrated sufficient capacity for inclusion of minority interests or interests of people with less (or no) capital to invest in the DAO.

Quadratic voting, such as that implemented by Democracy Earth, allows people to express strong preferences for specific issues in situations where there is equality of representation to begin with. However, when it comes to cryptocurrency and funding of DAOs, representation is always relative to the amount of money that someone donates, even in quadratic funding, and the funding is independent of the people who are affected by the voting and funding. For example, Black Girls Code recently raised funding on the Gitcoin grants platform through quadratic funding. The voters are the funders, not the black girls who will be affected by the grant. While there is nothing intrinsically wrong with that, it isn’t a form of democracy where those affected by a decision are those who make the decision. Similarly in the Colorado example of quadratic voting, by the way. The democratic representatives of the people participated in the quadratic voting; the people they represent did not.

One of the great failures of democracy is the disconnect between law-making and results being accomplished. Laws are implemented and continued for decades without review of whether their execution and implementation has accomplished the desired outcome; and when they do come under review, there often is no mechanism for repealing the law, but only to improve or adjust the execution of the law. DAO technology needs to include feedback mechanisms that will allow rapid adjustment when the measures are not met.

DAO technology has excelled in automated execution of decisions. For code changes, this is a complete process. Aragon and GovBlocks include mechanisms that allow code to be integrated automatically into the blockchain. However, this approach falls short when it comes to distribution of funds. Groups and individuals receive funds upon approval of their proposals, but none of the DAO systems to date include an accountability process. If the funds are misused or absconded with, there is no mechanism for holding the group accountable for the work. Recent work by the SEEDS project on Hypha DAO technology is developing a mechanism for escrow and then a release mechanism, which will increase accountability.

Accountability for more complex problems is even more difficult to track. For example, to improve the water quality of a river, it’s not enough to just execute a proposal; the water quality needs to be measured. It’s quite possible that the idea doesn’t prove itself in reality or that additional measures are required. Feedback loops should be developed to identify when decisions are incorrect, and adjustments made.