At the beginning of September 2019, Gemma Newlands presented the paper 'Pseudo-AI: The Workplace Implications of Ontological Obfuscation' at the Annual Meeting of the Society for Social Studies of Science (4S). The conference took place in New Orleans, USA, from the 4th to the 7th of September 2019. It convened leading STS scholars from across the globe and is the largest and most prestigious STS conference. This year's conference had the theme 'Innovations, Interruptions, Regenerations'. Gemma's paper was part of the panel 'Disturbances, Recreation of Labor: AI, Robots, Platforms, and Algorithms'. The abstract of her presentation is provided below:
One argument against widespread Keynesian ‘technological unemployment’ is the ongoing secret reliance on human labour in the execution of many ‘AI’ services. This phenomenon, referred to as ‘pseudo-AI’, involves a process of ontological obfuscation whereby technological deficiencies are bootstrapped through the use of human workers masquerading as AI. Although AI systems are ontogenetic, consistently upgrading in a process of extended-beta, pseudo-AI goes beyond the standard use of human workers to oversee the AI-training. Instead, human workers (in usually remote call-centres) either a) fulfil the tasks explicitly described as automatic/algorithmic and/or b) communicate to users pretending to be the AI in a process reminiscent of a reverse Turing Test.
In current discussions, pseudo-AI has been frequently conflated with the ‘Wizard of Oz technique’. However, aligning pseudo-AI with experimental simulations glosses over the significant implications which pseudo-AI has for labour recognition, business ethics, STS theory, and communication theory. The purpose of this paper, therefore, will be to explore the broader implications of pseudo-AI, while focusing on the implications for the invisible workforce who are not only labouring alongside AI systems in an increasingly posthuman workplace, but are in fact labouring in the place of AI systems in an ironic process where the humans have in effect stolen the AI’s job. This paper will also examine how workers can resist processes of reification and technological sublimation through attempts at ‘breaking’ the façade and asserting their own humanity to users and other external stakeholders.