AI based inventions mean humans will no longer understand emerging technologies, causes rogue AI.
Dornis 20 – Global Professor of Law at New York University School of Law, Chair of at Intellectual Property Law at Leibniz University, Professor of Law at Leuphana Law School, PhD in Law and Economics from Tuebingen University, LLM from Columbia University, J.S.M. from Stanford University.
Tim Dornis, “Artificial Intelligence and Innovation: The End of Patent Law as We Know It,” Yale Journal of Law & Technology, Fall,
Yet the effects of “artificial inventions” on the contents and, ultimately, on the theoretical understanding of the state of the art may prove detrimental—at least in the long run. Of course, we might expect the emergence of new types of inventions and discoveries, maybe so elaborate and unexpected and beyond human understanding that the future will become one big bang of knowledge.177 But we might also expect the opposite. After all, if the state of the art is to be increasingly shaped by artificial actors and artificial inventions, issues of content and quality will come to the surface. This does not primarily concern whether inventive AI acts illegally or unethically. It is the simple fact that inventions are artificial that is problematic, especially since our theoretical understanding of the state of the art will inevitably deteriorate over time.
The phenomenon of discovery in the sense of “first answers, then explanations” is well known. Acetylsalicylic acid—the major ingredient in modern-day aspirin—was used by shamans in early cultures as a remedy against fever and pain. 178 Its industrial manufacture started in the nineteenth century; a patent was granted in 1898. However, the substance’s actual, physiobiological effects on the human body were not described until more than seventy years later.179 This unexplained use of effects without a theoretically founded understanding has been described as an accumulation of “intellectual debts.”180 With an increase in AI-based inventions, debts of this kind will grow. As we have seen, algorithms work through brute-force computing, almost always in unexpected ways. They may find relationships in large chunks of data, but (at the moment) they will never produce a theoretical description or model of causalities. 181 Artificial inventions will thus yield results in the sense of unexplained discoveries—and making use of this kind of output will amass intellectual debts. 182 Even if the use of a single invention of this kind may not raise concerns at first, it is the combination of many such debts that carries a risk. Over time, algorithmic inventions will increasingly resort to prior algorithmic inventions—in other words, algorithms will invent “on the shoulders of” other algorithms. If we lose our understanding of theoretical relationships at the first level, how much more uncertainty will arise with regard to the artificial results that algorithms will create in the future?183
The consequences do not require a great deal of imagination. If we know only whether something works, but not why and how, the risks grow exponentially with each new wave of algorithmic invention. The more we rely on seemingly superior AI skills, the more difficult it becomes to explain foundations and make predictions. Incidental malfunctions of algorithmic aircraft control systems or so-called flash crashes in artificially managed stock markets are evidence that the algorithmization of modern life will become progressively uncontrollable.184 This intellectual artificialization will ultimately dilute our understanding of the world and how it works. Indeed, we will increasingly live in a dream-like state where things simply work and we cannot ask how or why. At least in the long run, we have to ask whether the patent system will still be able to function as a repository of generally accessible and, most importantly, actual knowledge and information.