Compeerism solves racist algorithms and data colonialism

Hannes, Professor of Geography at the University of West Georgia. “Taking Back the Internet.” From Capital to Commons: Exploring the Promise of a World Beyond Capitalism. University of Bristol. Pp 68-69. /LJS

Following from this, a strong compeerist case can thus be made for any commons-derived aggregated, open-sourced data to be placed under more proactive peer-property licensing regimes that limit how and by whom the data may be used. Enforcing such regimes for aggregated data, however, is even more challenging than what is involved with specific, added-value intellectual property. Not only is it difficult to contain and control access to this data once it has been made available to a large number of in-group entities, but proving that it has been utilized by an out-group actor in a specific way is often problematic. Nevertheless, in certain situations, where the misuse of protected data is believed to be especially egregious, targeted legal challenges could be pursued on the basis of the relevant peer-property stipulations in place. In the longer term, of course, as part of a broader economic shift to the commons, the compeerist aim will be to increasingly diminish the power of capital in all aspects of the economy, including in the area of data collection and use. Still, it is important to acknowledge that such a decapitalization will not automatically deliver us from the ills linked to the use of agglomerated data today. This is because the underlying technology, developed to assist in the functioning of capital-focused assemblages/apparatuses, could still be present, retaining a latent agency that runs counter to compeerist aims. Such agency is particularly evident in capital-sourced AI and algorithms, whose code is full of profit-seeking and class-based/racial biases. Hence, in line with abstract hacktivist principles, this code must be systematically expunged through meticulous rethinking and rewriting. Yet even then, we must ultimately accept that algorithms can never be fully objective or foolproof; being human constructions they will always be reflective of the prejudices and assumptions inherent in humans and society more broadly (Ntoutsi et al, 2020). Following from this, the enlightened use of digital networks and their trove of endless data will require nothing less than continuous critical, democratically informed oversight and engagement. Couldry and Mejias’s (2019) answer to this call is to involve the populace in an ongoing, citizen-based research endeavour that seeks to keep tabs on how digital data-capturing technology works in the real world. To execute such a project, an open database could be maintained that tracks a wide range of important algorithms and their functions, based on people’s submitted personal experiences. Ideally, then, where problems are identified, collaborative solutions could be found to restructure the hardware, programming, and the overall governance arrangements overseeing this technology to ensure that any bias or other forms of unfair targeting are purged (Stewart, 2020).