Objet Petit’AI and the ethics of infliction
A century ago, modernity dawned with sublime terror.
God was dead, and metamorphosis was everywhere. We dreamed of transfiguration, an ontological brightening that saw in the electrical quickening of machines our own bodies and way of being. Once again, we’re living in inspired times. Artificial general intelligence—not boring machine learning but the real shit: neural networks and affective computing and self-manifesting consciousness, just like in the stories—is a spectre burning, beguiling, at the threshold of Society 5.0.
Bombs, Autobahns, Bluetooth. Technology is infrastructure. As we build machines they build us back, and as technology probes deeper into life’s mysteries, splicing into questions of being, the right of a creator to implicate human experience and ontology is vague.
Unlike human rights, there is neither an epistemology nor a tradition of natural law contemplating the metacolonial claim of technologies like widespread general AI or, for that matter, interplanetary colonization and plastic bags. After all the destruction of modernity, we’re still looking for an ethics of infliction, and one of non-involvement as well.
But newness is exciting because technology appears altruistic and relieving of the past’s hindrances. Drone-filled skies and underground Hyperloops are easy to accept as inevitable because creators easily back their daydreams with power. That’s an issue for AI, because normalization is attainable in its early stages, when AI is what it eats. For the most part, the data we feed it determines how and what it learns, so projects to scale AI for societal functions—sentencing, for example—can magnify the subjective subtexts of data, and configure even subtle biases into algorithmic conclusions.
At this inflection point, a tiny group of venture-backed researcher-entrepreneurs embodying contemporary structures of privilege and power decide what’s important for AI, formalizing new values of data ownership, autonomy, and ideologies around risk and benefit.
Soon, opting out may be a privileged estrangement. Next month the “right to be forgotten” will be codified into EU law (GDPR Article 17—the right to erasure), compelling controllers to destroy personal data when users withdraw consent. While some argue that transformative technologies like AI and blockchain will be fundamentally incompatible at a moment of intense promise, the law is striking by establishing a right to not be personally implicated in digital technology.
The timing is crucial. AI has left the labs of its origin. From here on out it’s all about capital, baby. Like any capital, it will exploit the commons for its purpose, even as we assent through a widespread, permissive attitude towards technology which privileges corporate over individual rights, interprets privacy through the lens of free speech, and presumes the perpetual growth of individual choice.
Who’s in control here? This century brings its own sublime terror, this time in a computational theory of mind. Deep inside, existence is documented in bits. Somewhere even deeper, those bits are quantum impulses unbound by space and time.
Is this immortality? The daydream of AI is not the death of god but to become god, to transfigure the subjectivity of being into infinite formulations. Some have eternal faith in god; others in the sublation of consciousness itself. Others, who desire erasure, have faith in none of it.
God was dead, and metamorphosis was everywhere. We dreamed of transfiguration, an ontological brightening that saw in the electrical quickening of machines our own bodies and way of being. Once again, we’re living in inspired times. Artificial general intelligence—not boring machine learning but the real shit: neural networks and affective computing and self-manifesting consciousness, just like in the stories—is a spectre burning, beguiling, at the threshold of Society 5.0.
Bombs, Autobahns, Bluetooth. Technology is infrastructure. As we build machines they build us back, and as technology probes deeper into life’s mysteries, splicing into questions of being, the right of a creator to implicate human experience and ontology is vague.
Unlike human rights, there is neither an epistemology nor a tradition of natural law contemplating the metacolonial claim of technologies like widespread general AI or, for that matter, interplanetary colonization and plastic bags. After all the destruction of modernity, we’re still looking for an ethics of infliction, and one of non-involvement as well.
But newness is exciting because technology appears altruistic and relieving of the past’s hindrances. Drone-filled skies and underground Hyperloops are easy to accept as inevitable because creators easily back their daydreams with power. That’s an issue for AI, because normalization is attainable in its early stages, when AI is what it eats. For the most part, the data we feed it determines how and what it learns, so projects to scale AI for societal functions—sentencing, for example—can magnify the subjective subtexts of data, and configure even subtle biases into algorithmic conclusions.
At this inflection point, a tiny group of venture-backed researcher-entrepreneurs embodying contemporary structures of privilege and power decide what’s important for AI, formalizing new values of data ownership, autonomy, and ideologies around risk and benefit.
Soon, opting out may be a privileged estrangement. Next month the “right to be forgotten” will be codified into EU law (GDPR Article 17—the right to erasure), compelling controllers to destroy personal data when users withdraw consent. While some argue that transformative technologies like AI and blockchain will be fundamentally incompatible at a moment of intense promise, the law is striking by establishing a right to not be personally implicated in digital technology.
The timing is crucial. AI has left the labs of its origin. From here on out it’s all about capital, baby. Like any capital, it will exploit the commons for its purpose, even as we assent through a widespread, permissive attitude towards technology which privileges corporate over individual rights, interprets privacy through the lens of free speech, and presumes the perpetual growth of individual choice.
Who’s in control here? This century brings its own sublime terror, this time in a computational theory of mind. Deep inside, existence is documented in bits. Somewhere even deeper, those bits are quantum impulses unbound by space and time.
Is this immortality? The daydream of AI is not the death of god but to become god, to transfigure the subjectivity of being into infinite formulations. Some have eternal faith in god; others in the sublation of consciousness itself. Others, who desire erasure, have faith in none of it.