Intellectual Property and AI: The Ownership of Possibility

Those monitoring the deep learning related patent filings³⁹ occuring in the U.S. over the past few years have observed an inordinate amount of lawsuits regarding intellectual property for machine learning algorithms and heuristics.⁴⁰ This is obscenely controversial because these optimization and classification techniques are essentially mathematics wrapped up in applications. When properly understood, the implications of this should exceed concern and rapidly escalate into outright panic. This is, simply put, a legislative power play to stake out a monopoly on the future from the emerging vectorialist class.

Allow me to elaborate insofar as I am capable: by owning these types of connectionist systems (neural nets, etc.) these companies are able to make complex projections about consumer behavior, cultural vectors, market proclivities, transnational interests, and natural & biological systems that exceed anything we've ever encountered.

These companies are claiming to own statistical techniques which predate these companies’ existence. By obfuscating the underpinning mathematics with technical applications for patent law purposes these companies are alleging ownership on things that they didn't create. Owning these types of learning algorithms not only assures the imminent obsolescence of worker-ownership but it guarantees no dividends or returns for those that have created the value. This power lies in monopolizing intellectual property — patents and brands — and the means of reproducing their value — the vectors of communication. This transcends the ownership of the means of production into the means of possibility. This is what Wark means when they say Vectoralism.⁴¹



In the quadrant above we see the mapping of artificial intelligence’s ideological proclivities tending toward centralized power. I think this assessment is largely correct given the resource necessity and investment capital required to compute that much data. This rendition of the political compass axis meme was likely created with a sense of hyperbolic wit, but it ensnares an elemental veracity: ideologies are latent within these types of frameworks. I am not quite as polemical as Rao’s memefied compass and believe that it cannot be solely reduced to the authoritarian (“platform knows best”) vs. libertarian ("I own my data"). This is where we need to break open the differing layers of infrastructure ownership and discuss the merits of decentralized asynchronous back propagation, platform co-operatives, and unpack the data-as-labor discourse. I will address these subjects in chapter three of this essay and will elaborate upon them with some musical developments.

The subjects of this chapter’s compositions and theses are to draw attention to the risk of totalizing ownership that may happen over the course of a single generation. The capacity to surveil, collect, and compute the magnitude of data currently being amassed has the potential to end half of existing labor markets in a matter of a few decades.⁴² Data aggregation has modified the technical composition of capital and has produced changes in value composition. Automation combined with the rate of capital accumulation will inexorably produce unforeseen levels of unemployment assuring the constant presence of a “surplus population” or “reserve army of labor.” Without some of the interventions outlined in the third chapter, this engine of indifference runs the risk of stratifying a class disparity that the West hasn’t seen in millenia. The title Of Strings and Kings isn't an empty aesthetic parable but is invoked to reveal a stark convergence of imminent realities.

---