How Anonymity revolutionizes how we share ideas
We are building Lógos to change how we communicate in DAOs.
One of the challenges of online communication is how we can engage and share ideas freely. This is especially important when discussing DAO proposals, which effectively control how DAO treasuries are spent.
How DAOs are affected by Bias attacks
The problem with these types of discussions is that they are prone to bias attacks that are either:
- Covert: players attempt to sway opinions by relying on their reputation or connections while hiding it. E.g. Bribery
- Overt: players attempt to sway opinions by overtly campaigning and using their reputation
- Unconscious: bias happens subconsciously where we censor ourselves to either protect our relationships or our image
In a nutshell, how we are seen, or how we see others, has a significant effect on how ideas are evaluated and expressed. This is a well known problem in academia that is solved by double-blinding experiments and papers.
Additionally, it is compounded where in a setting where social ties matter, where allies don’t want to put existing relationships at risk or enemies are unable to agree with each other due to historic differences.
Does Anonymity really improve the conversation?
We have been iterating on Lógos for the last couple of months. Logos currently runs on IPFS and Ethereum, making it an unstoppable and totally anonymous platform. However, we needed to prove that anonymity does remove bias and improve DAO conversations.
A tale of two groups: One Anonymous and one Pseudonymous
In order to test its effectiveness we created an experiment with two groups
- Anonymous — with no knowledge of who they were talking to (experiment group)
- Pseudonymous — who knew each other on discord (control group)
We gave each group a controversial topic to discuss, with the pseudonymous one knowing who the proposer was, and also knowing each other.
We had each group discuss their thoughts on the topic and analyzed the discussions, and collected subjective feedback from participants, summarized below.
Analyzing subjective feedback from our users quantitatively
This indicates a balance in the random selection of participants
This indicates a general satisfaction in both Anonymous and Pseudonymous settings.
This tells us that subjectively, users felt the risk of influence that comes from knowing who the proposer is.
Although both groups felt the discussion quality was high in both cases, the difference made by anonymity was palpable in terms of evaluating the proposal.
This is of great import to DAOs wanting to have proposals evaluated by merit instead of influence.
Qualitative analysis from participant feedback
These are a sample of the feedback we received from different participants in the control group.
On trusting the proposer:
“Yeah, because I trust he’s a professor. Of Course, I read it carefully what he implied. it affected my ideas about it you know, because I respect the professor. So yeah of course it affected my ideas.”
On being able to freely express themselves:
“Because I’m in a public space and for example I can’t use certain humor or F words, like chatting with my friends…. I felt that I needed to be cultural in my wording so as not to offend anyone and at the same time express my real opinion.”
In the context of self-expression as a DAO collaborator
“if we were talking about a specific startup or project, and my opinion could be perceived as the opinion of the company — I would be more limited in my statements and would prefer anonymity for an honest opinion”
The Risk of Anonymity
Users also expressed concern of the potential for increased spam in an anonymous setting.
“Well, the cons, not at our wonderful meeting, but potentially — I think that anonymity attracts people who will be rude to other users, counting on impunity”
This is an issue we believe can be solved through staked reputation — a feature we are currently working on.
Summary of subjective analysis from participants
In summary, the following insights were gleaned from feedback from our users:
- Users felt that knowing who the proposer was did have an effect on how they evaluated the topic. (e.g. Expert vs non expert)
- Users felt less barriers to engage in the conversation as well as to put their honest feedback in the conversation.
- Users felt Social relationships and Company affiliations affect how they might respond to certain topics, depending on the context. For example, concern about opinions being perceived as the opinion of the company regardless of the intent.
- Users felt anonymity helped to focus on the merits of the idea rather than the identity.
Using AI assisted analysis to gauge sentiment
We used AI tools to analyze the conversations for sentiment and classify it into different categories, such as:
- Diversity of Opinions
- Emotionality
- Consensus
While interviewing users gave us their subjective feeling on how the discussion in anonymous and pseudonymous settings, passing their conversations through an analyzer also gave us objective feedback on the discussion styles.
After evaluating several alternatives, ChatGPT was selected for its versatility in classifying the conversation into multiple categories.
ChatGPT’s Analysis of the Anonymous and Control Group
Summary of Sentiment Analysis
Overall, the anonymous group had a more diverse emotionality and discussion styles with an increased polarity of ideas, in comparison to the control.
In order to discuss an idea fully, alternatives need to be explored by minimizing the risk of group-think. This creates a “market-place of ideas”, where the most competitive ones rise to the top.
However, reducing group-think can also make consensus more difficult, which is also important for the DAO.
We have some great ideas in the pipeline to enable a gain in the advantages of anonymity while also:
- Minimizing spam
- Enabling easier consensus
- Replacing influence with expertise
These ideas will be explained more as our product matures through feedback-loops with our users, and also with the release of our whitepaper.
Join us in our journey — help us improve how DAOs make decisions through Anonymity.
We are in alpha mode and need testers to try our current ideas and also come up with new ones. Visit our discord and ping @skyfoxx to get the tester role