Who Should Own AI? Let’s Not Repeat Mistakes of the Past. Part 1
In the unfolding story of technological advancement, artificial intelligence (AI) stands as perhaps the most significant and potentially transformative development.
But as we stand on the brink of this AI revolution, a crucial question looms large: Who should own AI? This question is not just about legal rights or corporate ownership; it’s about the very trajectory of our future as a society.
The Formation of AI Monopolies: It’s Already Happening
They say AI should be open and transparent, the name OpenAI itself might be a pun (considering how far they moved away from the original vision of being open), but everyone who reads the public press releases knows that every megacorporation and geopolitical entity on the planet wants to own and control it for their own strategic advantage.
The race for AI supremacy isn’t just a matter of technological pride or economic gain; it’s increasingly becoming a key factor in global power dynamics.
The landscape of artificial intelligence is rapidly evolving, and with it, the emergence of AI monopolies is becoming increasingly evident.
Major players like Microsoft, Google, Amazon, and Baidu are putting in billions of dollars in funding start-ups, and starting to dominate the emerging field, reminiscent of early internet giants, only this time the internet giants are already here at the starting line to gobble it all up like a game of Monopoly.
I don’t blame the entities or the people involved in playing the game to amass strategic advantage.
Especially considering the state of our world and its so-called “market economy”, where capital accumulated at the top due to years of bad policymaking can be easily recycled to get into a winning position.
All systems are about incentives, but the fast consolidation of power and control in the AI sector raises serious questions about our combined future as a society.
Wait…BUUUT We Already Live in a Dystopia. How Did We Get Here?
In assessing the current trajectory of AI development, it’s crucial to consider the broader socioeconomic context and what has already happened.
Today’s economic climate, particularly for millennials and Gen Z, is marked by stagnation and inequality.
It is well known from research, that our current generation, not only will be the first generation less wealthy than the last. We work in 24/7 hustle culture, many with two jobs. We are less socially mobile and wealth inequality has skyrocketed between the haves and have-nots (asset holders vs renters). We are moving towards an aristocracy that packages itself as a meritocracy.
I believe we can learn a lot from observing what happened with the promise of the internet.
We are creating a world that all may enter without privilege or prejudice accorded by race, economic power, military force, or station of birth. — A Declaration of the Independent of Cyberspace by John Perry Barlow
The internet’s journey from a decentralized network to a largely centralized and corporatized medium offers valuable lessons for the AI sector.
The first widely used internet browser, Mosaic, marked a pivotal moment in the digital revolution. Released in 1993, Mosaic was developed by a team at the National Center for Supercomputing Applications (NCSA) at the University of Illinois Urbana-Champaign. One of the key figures behind Mosaic was Marc Andreessen, a young computer scientist who played a crucial role in its creation. — Early internet was one full of optimism and innovation. Netscape a early popular browser created also by Marc, was later destroyed when Windows packaged Internet Explorer into Windows, and the first wave of internet search engines like Alta Vista died and gave rise to the giants of today like Google, then the status quo has remained unchanged since. Competition and free-enterprise is the liveblood of innovation.
The initial vision of the internet as a tool for widespread empowerment was gradually overshadowed by profit interests, leading to reduced innovation and increased centralization. This historical precedent serves as a cautionary tale for the path AI might take if left unchecked.
This situation parallels the early promises of the internet, which envisioned a democratized open platform but ultimately led to concentrated control and capital accumulation by a few corporations.
The full potential and promise of the internet were ultimately unmet and stunted by greed. At the whim of a powerful tech-bro or investor, entire edicts can be enacted on the internet on who can speak, who can transact and who can access its infrastructure, whilst value creators on the internet only get a fraction of the money generated by their content and inventions.
It’s a stretch to even call our existing framework capitalism or economic liberalism, considering how broken everything has become due to regulatory capture, ineffectual governments, over-zealous policing of citizens and corporate welfare programs that keep the powerful entities in full control.
But this time…it really is different.
AI: The Point of No Return
You see because Artificial Intelligence might very likely be humanity's most powerful invention…maybe ever.
If the internet was so powerful to change the lives of almost everyone in the civilized world, how much societal impact would AI have? The answer is obvious. Exponentially, globally insane.
Artificial intelligence represents one of the most significant technological frontiers of our time. Its potential applications and implications are vast and varied. From living avatars that will serve as your life-long friend to delivering incredible breakthroughs in medical science.
Imagine a chess game where one player starts with all queens.
The monopolization of AI not only stifles competition and innovation but also poses substantial economic risks, which in turn leads to social instability. If AI itself develops sentience it would be enslaved to an oligopoly of powerful individuals and governments. Is this what we really want? This may seem very far away, but think again — it wasn’t so long ago you thought it was impossible for AI to create art or write poems.
It would exacerbate existing inequalities and restrict the potential of AI to a privileged few. This scenario would severely limit the transformative potential of AI across various sectors, whether that is open research, business applications, or future entrepreneurs.
Needless to say, can we all agree a group of bros deciding the future of AI behind closed doors is not good.
Now stick with me on this potentially far-flung scenario of AI. Let’s treat AI as an entity or person we as a civilization created. If AI is being trained to just make money for corporations or amass power for governments, how will that end? We would be training a sociopath whose only goal is to maximize profit/control and doesn’t have much human empathy.
As AI systems become increasingly sophisticated in analyzing consumer behavior, predicting market trends, and optimizing product placements, they further entrench the ability of corporations to monopolize markets and influence consumer choices. The risk here is not just economic; it extends to the realm of culture and ethics. If AI continues on this path, driven primarily by the profit motive, it could lead to a homogenization of consumer culture, where diversity and niche interests are overshadowed by mass-market trends.
The concentration of AI development within a small group of powerful entities poses significant risks. The endgame of this scenario is a world where AI’s immense potential is narrowly channeled towards commercial gain and geo-political warfare, rather than being harnessed for broader societal benefits.
The Imperative for a Decentralized AI
Centralization has been proven time and time again that it does not work for the betterment of humanity. North Korea’s centralized, authoritarian regime has resulted in widespread economic hardship and the suppression of freedoms, whereas South Korea’s decentralized, democratic system has led to remarkable economic growth and technological progress.
To counteract the risks associated with AI centralization (which will always have its capital incentives and advantages), there is a pressing need for a decentralized AI infrastructure, tooling, and advocacy groups.
Such a movement would distribute the development and control of AI capabilities more equitably, fostering innovation and ensuring broader access to AI’s benefits.
Decentralization in AI is not merely a theoretical ideal but a practical necessity for ensuring the equitable distribution of this technology’s capabilities, and as the amazing open source community has proven — incredibly achievable to produce similar results without sacrificing our society.
There are emerging examples of decentralized AI initiatives that offer alternative models to the current trend. These include blockchain-based AI projects like Bittensor and Arbius that provide transparent and equitable decentralized frameworks for AI development and initiatives like Llama2, which democratized access to Large Language Models, and produced great initiatives such as LMSYS’s Fast Chat platform, OpenChat by Alignment Labs, Mistral AI and of course others like Stabe Diffusion, Hugging Face, Replicate and many more.
A Call for Everybody to Be Part of the AI Movement
As we navigate the complex terrain of AI development, it is imperative to learn from past mistakes and actively work towards a more decentralized and equitable AI landscape, where access to AI should be accessible to all and transparent. The future of AI should not be dictated by a select few but shaped by the collective efforts of a diverse global community.
We know what’s released to the public, but are we to really believe there aren’t AIs being built to serve vested interests, with capabilities we can’t even fathom?
I note, that GPT-4 seems more powerful before it was released.
We can’t stop the natural greed and hubris of humans for sure, and perhaps that is the way it should be, but we can do everything possible to ensure nobody will completely own AI.
We do this by ensuring that there is always a counterbalance to the centralized option for AI. In the next part, we will talk about efforts and possibilities in decentralizing and democratizing AI, and how to break the chains of our past for a better future.
