From DAOs to AI DAOs. Blockchain + AI driven organizations.
In one of my recent articles I reviewed Andrew NG’s “The State of Artificial Intelligence” speech. One of my points in that article following Mr. NG’s statements is that being an AI company requires to be a digital company in the first place.
Although I firmly think AI will permeate every challenge of every organization, little was mentioned in my article about what it could imply to management or decision making structures. This article will explore in a very high level how AI, when combined with other technologies like blockchain, could change how organizations work and decide, thus disrupting current management structures.
First of all, and before getting to the role of AI, let’s analyze the concept of the DAO, which I think constitutes a strong base for decentralized and automated management in which AI can play a very important role.
What is a DAO? DAO stands for Decentralized Autonomous Organization. DAOs rely on Smart Contracts to encode its rules and in a blockchain to keep track of how are they applied. Putting it simple, I especially liked the definition found in this article:
“An organization that runs autonomously, in a decentralized manner, that functions without the need for centralized parties to make decisions for the organization to grow, to be profitable, or *physically* exist.”
https://www.youtube.com/watch?v=Pyi8-qm02hs
There are already some real examples of this kind of organization like The DAO company, Digix.global or Terra0 (more on this one later).
Leaving ICOs and their controversies aside, it is clear that DAOs can provide a series of advantages both from a company internal and external point of view:
- Bureaucracy can be eliminated as far as clear rules are defined in Smart Contracts.
- As decisions follow the rules defined in Smart Contracts, hierarchical structures do not have negative effects on predetermined expected behaviours.
- As an investor/ stakeholder, and as long as you are in possession of some DAO tokens (kind of its shares), you are entitled to participate in the company decisions in a transparent way, thanks to the blockchain and its capabilities.
So far so good. Using DAOs would seem a good way we can create companies ruled autonomously and in a transparent way. But before describing how AI can improve these DAOs, let’s see some of the problems with them. For that purpose, we are going to use one of the real implementations we have seen before, The DAO company.
The DAO Company, born as a decentralized venture capital fund, was created over the Ethereum blockchain and allowed investors to decide on investment proposals based on ideas submitted by contractors and curated by a group of volunteers. It claimed to be completely transparent and eliminated the need of fund managers or any other kind of management structure.
Unfortunately, The DAO Company ran into a problem linked to one of blockchain’s main features: immutability. Thanks to a vulnerability in the DAO’s code, the organization was compromised and one third of its funds derived to an external account. Actually, this episode caused one of the biggest controversies in the Ethereum community, leading to a hard fork of the blockchain in order to allow restoring all funds to the original contract.
Could AI help avoiding these situations? As far as the blockchain is very difficult to hack nowadays, applications built on top (the DAOs) are not guaranteed the same level of security. Through the use of anomaly detection algorithms, and due to the fixed structure of blockchains, AI could act as an early detection system for bugs like the one described above. Nevertheless, this would not be enough to avoid the core problem, due to the immutability of the contracts.
So, how is mixing DAOs and AI helping the future of organizations? Trent McConaghy, from BigChainDB has some answers for that. You can find what he defines as the three paths to get to the AI DAOs in this article. Long story short, if we were able to successfully apply AI to DAOs, we could be creating the first AI-managed autonomous and decentralized company. What does this mean both from an internal and external perspective to how companies are ran?
- As blockchain encourages data sharing by design (thanks to immutability, auditing), better data relevant to our business would be available to our algorithms, eliminating silos.
- AI would be able to use that relevant data producing decisions with a CEO reach (cross-department/ strategic vision) but with the same fine-grained information that any department manager would use, thus eliminating proxys.
Ideally, we would be creating what we could call an AI C-Level, in which different AIs could look for different parameters (e.g. economics, branding, human resources, etc.) and take the best possible decisions both for the employees and shareholders.
The following quadrant by Vitalik Buterin describes how applying automation at the center and edges of a company makes us evolve from the “boring old organization” to the “holy grail” of the AI company.
Remember that we talked about Terra0 before in this article. In its concept paper, specifically on the technical implementation chapter, they describe different levels of complexity for the AI DAO, being the highest level one in which “ The smart contract is no longer distinguishable from a completely developed artificial intelligence”.
This really nice article describes different examples of “Decentralized Intelligent Companies”. Although not that many for that moment, it is clear that this kind of AI + blockchain based start-ups are gaining momentum.
Finally, I would like to point out that this kind of approach mixing definitely brings together the best of two disruptive technologies like blockchain and AI, but it also brings (and augments) some of its problems into the table:
- How do we make this kind of autonomous companies compliant with GDPR or other data protection regulations? As we saw before, immutability of the blockchain is a nice asset but also implies some problems, e.g. we cannot forget data.
- Also, in terms of legal compliance and GDPR, AI introduces another problem: how are this companies going to justify their decisions if necessary?
- Even if we push decentralization to its limits and no regulation at all is required: how we ensure the effects of completely autonomous AI-based companies on our social structures?
Definitely DAOs and AIs will constitute a green field for technical experimentation and business innovation during the next years, given regulation evolves accordingly. Legal aspects as personality or jurisdiction (amongst others) and how are they considered by law enforcers will have a great impact in the development of AI and DAOs. Some controversy has already been raised around recent regulatory proposals, considering assigning legal personality to DAOs, or even applying the “nearest person principle” to them.
How do you think DAOs and AI will evolve together in the future?
