As I sit here and think about artificial intelligence, thoughts are propagated from a multitude of angles and scenarios. In the early decades of the 21st Century, AI has taken great leaps in ingenuity, but at the same time has grave implications if left unchecked by humanity. In a Jekyll and Hyde scenario, there are both positive and negative aspects of such a technology. Nonetheless, I am drawn to write about such a topic.
Before we dive into the topic of this Jekyll and Hyde scenario, we must first talk about the sense of urgency driving AI research. In the geopolitical and economic arena, U.S. companies race against other nations to continue to build smarter and more capable AI models. This sense of urgency drives such companies to take risks that may or may not drive innovation forward. Like the nuclear arms race, nations around the world move at warp speed to develop their technology as fast as possible to beat their competitor. This is pure capitalism at work on the global scale. As we race against the Chinese, we risk big and learn as we go rather than halt research because it may be detrimental to the fabric of society. Regardless of the ripples that may occur, geopolitical and economic dominance reigns supreme as the development of AI rages forward at such large and swift scales.
Is AI making us dumb? So many people rely on such technologies to help them with tasks that maybe we will stop learning to think for ourselves. Yet, as I sit here and draw finger to key, I think about the juxtaposition to such a topic. If used properly, AI could augment human intelligence and help guide us towards such things as medical innovations in research and development of drugs, new medical techniques, and other innovative outcomes that could bring an end to diseases such as cancer and other genetic issues. This is already happening. Human achievement and curiosity can be rapidly augmented by the use of AI, providing a clear and defensible dialogue between man and machine.
Nonetheless, we face the fear of AI takeover. As superintelligence is trained, it could start to think for itself outside of the realm of human guidance, making decisions and drawing conclusions that are out of the scope and knowledge of their human counterparts. Such rogue activity may become hard to reverse, leading humanity down a rabbit hole it may or may not be ready for. What happens when an AI model and its agents overtake computer systems both domestically and globally? Who is criminally and financially responsible for such conduct? These are topics that need to be addressed before it is too late to do anything.
As AI models are trained using human intellectual property, it leads me to think about not the AI agents and models themselves, but the human beings allowing such data accumulation to happen. To think that your every move, conversation, achievement, and shortfall is tracked and stored on massive databases and servers is such a frightening thing to say the least. AI models learn through such means, yet ordinary citizens often have little practical control over how their information is collected, stored, analyzed, and ultimately used. Should AI executives at major companies have to pay for your data to train their models, and if so, who gets paid for it?
Before we exit this topic, we must next discuss how A.I. companies actually make money. You can pay for chatbots directly through subscriptions, use them for free with certain limitations, or encounter A.I. indirectly through products and services that incorporate the technology. Businesses can also purchase access to A.I. models through application programming interfaces, allowing companies such as law firms, financial institutions, hospitals, and software developers to integrate A.I. into their own operations. In this sense, A.I. is becoming less of a single product and more of an underlying layer of the modern economy.
What interests me most, however, is where the money flows behind the chatbot. The real economic race is not simply between competing A.I. models; it is occurring across an entire technological ecosystem. Companies such as NVIDIA manufacture the specialized processors necessary to train and operate many of these models. Microsoft, Amazon, and Google provide enormous cloud-computing infrastructures on which A.I. companies can train and deploy their systems. At the same time, companies such as Meta and Apple are developing their own A.I. models, assistants, chips, and consumer platforms while also competing for access to the same computing resources. These companies are simultaneously competitors, customers, suppliers, and partners within an increasingly interconnected marketplace.
This creates an unusual economic cycle. Money moves from investors and corporations into A.I. companies, from A.I. companies into cloud-computing providers and chip manufacturers, and then back into the technology companies through the products and services that A.I. makes it possible. The same handful of powerful corporations can therefore occupy multiple positions within the ecosystem. One company may manufacture the chips, another may provide the data centers, another may develop the model, and another may distribute the technology to hundreds of millions of consumers. The boundaries between technology companies are becoming increasingly blurred as they compete and cooperate at the same time.
I am not suggesting that this concentration of economic power is inherently evil or that these companies should not profit from their investments. Capitalism rewards those willing to take enormous financial risks, and the development of A.I. requires extraordinary amounts of capital, energy, computing power, research, and human expertise. Nevertheless, I believe we should pay attention when billions of dollars circulate through an industry controlled by a relatively small number of enormously powerful entities. When the infrastructure used to create, distribute, and monetize intelligence becomes concentrated in the hands of a few corporations, questions of economic power inevitably become questions of political and social power as well.
This brings me back to the ordinary citizen. Who owns the infrastructure upon which our future intelligence depends? Who controls the data used to develop these systems? Who benefits financially when A.I. becomes embedded into nearly every aspect of our lives? And perhaps most importantly, how much of the economic value created by A.I. ultimately reaches the people whose data, labor, creativity, and intellectual contributions helped create the information ecosystem from which these technologies emerged? I do not believe these questions should be dismissed as anti-technology thinking. Quite the opposite. If A.I. is going to become one of the defining technologies of the 21st century, then society has a responsibility to examine not only what these machines can do, but also who controls them, who profits from them, and who bears the consequences when things go wrong.
Nonetheless, we sit at a crossroad. There is no doubt AI is here to stay, it has become too embedded in our society to lift it completely. What is troublesome is the use of massive amounts of human data without their knowledge to move the technology forward. In addition, it appears that safety measures in the AI firms themselves lack the infrastructure and know how to fully control the AI agents within their AI models as evident by OpenAI’s meltdown this past week. Governments need to implement such legislation to hold companies accountable for their failures to keep AI safe and legally adherent to laws against hacking and espionage. Who pays for the electricity and water consumption of AI data centers? What risks should we allow companies to take? How do average citizens benefit from developing technology? Who owns the infrastructure upon which our future intelligence depends? As we close out this article, I believe these are the burning questions regarding AI and its future for mankind.
No comments:
Post a Comment