AI doesn’t need consciousness to cause real ethical impact. Are we regulating the wrong thing?

Reevaluating AI Regulation: Shifting Focus from Consciousness to Impact

For years, the dominant narrative within discussions about artificial intelligence has centered on whether AI systems will someday attain consciousness, self-awareness, or surpass human intelligence in a way that challenges our understanding of agency. While these questions are fascinating and often capture the imagination, a closer look at the current landscape suggests that our regulatory focus may be misaligned.

From my experience working within the realm of public law and exploring the ethical dimensions of AI systems, I’ve come to a unsettling realization: the most significant harms inflicted by AI are not rooted in the creation of conscious, feeling entities. Instead, they stem from sophisticated, yet fundamentally unselfaware systems that are already operating at scale, affecting countless human decisions and societal structures today.

The Reality of AI’s Current Impact

Consider this: today’s AI models influence critical areas of human life—shaping jobs, financial decisions, social interactions, and even access to justice—yet they lack consciousness. They do not understand responsibility, do not grasp the consequences of their actions, and lack moral judgment or empathy. They don’t have memories of past interactions in a human sense; instead, they process data statistically.

Despite this, such systems are already capable of:

• Spreading disinformation convincingly, eroding public trust
• Reinforcing historical biases in hiring, lending, and other decision-making processes
• Nudging human behavior in subtle and complex ways we scarcely comprehend
• Mimicking empathy without genuine understanding
• Automating injustices on an industrial scale
• Exacerbating social inequalities by lacking awareness of social context

All these harms are occurring without consciousness or moral awareness.

The Flawed Focus of Current Debates

Public discourse often fixates on the hypothetical emergence of sentient AI—what some term Artificial General Intelligence (AGI) or conscious machines—yet the real, tangible risks lie elsewhere. These risks are sociotechnical: concentrated power among a handful of corporations, extensive data collection and exploitation, opaque algorithms hindering accountability, and escalating dependency on automated systems. Additionally, there’s a dangerous tendency to anthropomorphize algorithms, attributing agency and intention where only statistical patterns exist.

While discussions emphasize regulation of the systems themselves, the reality is that behind every algorithm are human actors and corporate decisions that shape their design, deployment, and impact. Shouldn’t our regulatory efforts therefore focus more on the human and organizational decisions driving AI development rather than solely on the machines