‘Civilization’ has always been a precarious affair. Werner Herzog described it as a layer of thin ice over a sea of darkness and chaos.

It is also a concept that has been widely abused, being used as a hypocritical justification for such uncivilized acts as enslavement, wars, and colonisation. In the process, it has earned an understandable — though undeserved — negative reputation; but its capacity to be an agent of chaos itself is notable.

The ancient Romans, through whom the word entered the English language, understood it as the manner and form of human association befitting a ‘city’ or civitas — and took it to include the attitudes and conduct of people calling themselves ‘citizens’ (cives), and to encompass the institutions, ethical norms, and political framework that made it all possible and allowed it to perpetuate itself.

The benefits of such an arrangement are that it allows for the mutual, reinforcing contributions of individuals to be concentrated in a place, and for a span of time. If implemented well, it creates conditions permitting human beings to reach the fullest creative and intellectual potential and material prosperity that circumstances may afford.

Nevertheless, the order this arrangement confers on life’s affairs is artificial, not natural. The price we pay is the struggle to preserve it and pass it on to future generations. This takes not only consistent and concerted effort but is continuously challenged by the very ideas, innovations, and technologies that it allows to come forth.

The ‘layer of ice’ is indeed thin, and today we must accept that it is not only threatened by its own fragility but by its melting point as well. The mutual contributions of individuals to our collective knowledge have created the vast corpus of written text on which the latest generative AI (GAI) systems have been successfully trained to mimic our thinking and emulate our speech. The accumulated technologies — and, I might add, the philosophies — of the web that link, sift, and distribute this information securely and transnationally have produced blockchain.

The rise of generative AI (GAI) — exemplified by ChatGPT — and blockchain typify the disruptive technology challenges to our current age. Both threaten to replace jobs; GAI by its ability to simulate human reasoning and communicate through natural language, blockchain through its capacity to replace a vast economy of transactional intermediaries. Both can potentially facilitate cybercrime; GAI by generating automated cyber-attacks, phishing emails, scamming campaigns, deep fakes, and social engineering; blockchain through the application of cryptocurrencies to money laundering and other illicit transactions, facilitating ransomware payments, illegal marketplaces, and so on. AI has managed to distinguish itself as the latest addition to a long list of perceived existential threats of our own creation.

The perceived threats from these technologies stem, in part, from their moral ambiguity — they are power tools that may be put to any purpose. In the case of GAIs, a further concern is that the technology is still highly centralized, and many organizations that have built them have dubious reputations for managing personal and private information responsibly. There is an opportunity, however, to change this if we are bold enough to pursue it.

Decentralization is a broad trend across industries. Computers transitioned from centralized mainframes to personal computers; cloud computing — which represented a brief reversion to a form of centralization — is now being diffused in the form of ‘edge computing’ where servers are positioned closer to where their computing power is needed. Likewise, energy is becoming increasingly distributed and decentralized as utility power station capacity is redistributed across rooftop solar and renewable generation; and the batteries used to store this surplus are distributed across whichever electric vehicles happen to be connected to the grid at a given time. Adaptive manufacturing (3D printing) is another example of how highly centralized processes like manufacturing are distributing themselves throughout the economy. The decentralization of generative AI can be next, and the role of blockchain will be to enable it.

Driving this will be a wealth of privately-kept data and knowledge residing on institutional and personal computing systems. Enabling it will be the fact that AI can be trained in a distributed manner through methods like federated learning. Federated learning involves training a common AI model piecemeal on individual and institutional data. Because the resulting model is a neural network — a ‘black box’ — the data it was trained on remains private. A global model can be assembled through these individual pieces, and the results shared between all participants. Blockchain-based mechanisms for decentralized storage, the provenance of model training, and incentive payments to training participants will form the infrastructure that will make this possible.

With the decentralization of AI will also come the decentralization of its training; and with this will come the recognition that we, individually, are responsible for what we help create. What we bring forth will be a reflection of us, and the value it generates for this next chapter of civilization will be a function of the value that we contribute to it and the values that we instill.

This is an empowering prospect that should inspire us to seize the opportunity and motivate us to learn all we can about these technologies to shepherd their advancement while fully participating in the rewards.  Our future progress shall continue to hinge not just on the technologies we create, but on the wisdom with which we deploy them.

Dr. Peter D. Finn, PhD, NUS and King’s College London (KCL), Blockchain and DX Trainer at SMU, and Chief Instructor of Vertical Institute’s Blockchain & Cryptocurrency Bootcamp.

TNGlobal INSIDER publishes contributions relevant to entrepreneurship and innovation. You may submit your own original or published contributions subject to editorial discretion.

The future of AI-assisted development: A deep dive into GitHub Copilot X and GPT-4 with GitHub’s Shuyin Zhao [Q&A]