Balancing AI Innovation and Decentralization: The Case for Open Source and Blockchain Integration

Risks of AI Centralisation

In November 2022, ChatGPT first hit the scene. It became immensely popular and quickly became viral, garnering one million users in its first week after launching. Since then, we have seen ChatGPT continue to improve and gain popularity, alongside other AI models developed by companies such as OpenAI, Meta, Microsoft and Google. As AI is continuing to become more and more powerful, it is gaining hugely widespread use, as well as becoming increasingly less accessible to start-ups, due to its accelerating demands for energy and huge data centers. This means that this new, and immensely powerful, technology is finding itself concentrated and centralised in a few major companies. In a world becoming increasingly involved and reliant on artificial intelligence, the danger in this level of centralisation is clear.

Throughout history, we can see a plethora of instances where immense power has been concentrated in the hands of few, not least in the domains of militaries and dictatorships. The greatest and most relevant of these instances was the Manhattan project in the ‘40s, where a new and immensely powerful technology was centralised in the hands of a few war-focused US militarians. What was the consequence of that sort of centralisation? There were over 200,000 innocent lives lost, probably unnecessarily and certainly undemocratically. This was an undeniably terrible historical event, which we by no means want to relive – yet we are seeing a not so distinct situation today. Although not as immediately detrimental, ownership and development of artificial intelligence (and other similar emerging technologies) is being further focused in the hands of a few corporate leaders, with its impact and potential devastation up to them.

When centralisation is developed to this great an extent, large corporate leaders cannot be expected to exercise their power and authority beyond the interests of themselves and their enterprise. The greatest form of governance is that which is led by the powerful and benevolent monarch – yet there is no benevolence here, only unhealthy power. In the same way, we cannot expect governments to step in, with their all-loving spirit, to democratise and decentralise this concentrated power. Their interest cannot be conveyed by the mantra “of the people, by the people, for the people”, but rather in Orwell’s brilliant line, “the object of power is power”. Regardless, it is becoming increasingly paramount that democratisation and decentralisation occur – and maybe actually by the people.

A Case for Decentralisation

It is clear that we are existing in a world of ever accelerating technological prowess, and that is not something to be considered lightly. We have evolved much more since the days of rural and relatively disconnected villages, focused on farming and craftsmanship. We have evolved into a dramatically and globally interconnected world, with much more organised and powerful systems, institutions and technologies. You can no longer take your horse, fill his saddlebags up with your possessions and erase yourself from the intelligence and world of your former city. You can no longer carelessly declare war with your neighbouring regions without the risk of utter catastrophe. So it must be evident that to permit someone, or some institution, to be given power and control, you would surely trust them deeply and confidently. But do you? Do you trust these grand institutions and figures of power? Or is that trust replaced by a love of convenience, of laziness, of ignorance?

In our modern world, the need is rapidly increasing for decentralisation and democratisation across a plethora of domains, be it financial, or legal, or technological. And in those situations where there must be a prominent dose of centralisation and strong authority, such as in governance or leadership, there must be clarity and transparency. Transparency. Decentralisation. Liberty. These are the three necessary pillars of our modern society. The encouragement and adoption of these values, as a guide for technologies and organisations, is paramount. Without these, we can never be assured freedom, confidence and accessibility, and surely we value these. In a world of accelerating interconnectedness and technological prowess, it is vital that we encourage transparency and decentralisation.

Open, Collaborative and Transparent Development of AI

So how can this be implemented in the development of artificial intelligence, and where can we see its emergence?

Our goal is transparency and decentralisation, so there are two key domains that demand our investigation, namely open source development and blockchain infrastructure. Open-source code refers to code that is openly accessible rather than restricted to use and viewing from within the company or organisation that develops it. A common objection to open sourcing is the openness of vulnerabilities to hackers with malicious intentions, yet we have seen time and time again, with open-source software such as Linux and Android, that open-sourcing builds communities that will resolve vulnerabilities before taking advantage of them. In terms of artificial intelligence, we are already seeing this concept being implemented by the InstructLab team working from IBM and Red Hat. In essence, InstructLab is an open source project that “allows users to interact with, train, and fine-tune LLMs”¹ based on community-developed databases of knowledge and skills. This project shows great promise, and may be a major leader in the space of transparent and openly accessible AI.

Beyond open source AI development, another highly promising means of AI development and usage is with blockchain technology. Blockchains are transparent, private, openly accessible and, most notably, not centralised in a single entity. Although one individual or organisation may develop the software, once deployed, it is open and existing on the computers of everyone running the program. This is the best bet away from the concentration of power and AI technologies in the hands of a small and powerful few. The alternative to this centralised concentration is already being innovated, most notably by the DAIA (Decentralised Artificial Intelligence Alliance). The DAIA is a coalition of organisations and developers working on the development of decentralised AI, who are focused on creating transparent, open and collaborative artificial intelligence. One of its founding members, SingularityNET, allows organisations and users to collaborate on, share and monetise AI using blockchain technology, with their mission of democratising advanced AI technology. Overall, the integration of open source and blockchain infrastructure in the development of artificial intelligence is vital, and must be continuously encouraged and invested in.

How do you see it? Comment your take👇

If you enjoyed this article or are interested to learn more, you may be interested in some of the following articles on decentralisation and artificial intelligence: