Free Speech and Artificial Intelligence

URL SCAN: Free Speech and Artificial Intelligence
FIRST LINE: # Computer Science > Computers and Society

The Dissection

The paper converts an AI power problem into a rights-allocation problem. It correctly identifies recommendation systems as controllers of speech visibility and conversational AI as a challenge to conventional ideas of speakers, corporate speech, and recipients. But it explicitly stops at raising questions. It maps the constitutional surface while leaving the ownership and control layer largely untouched.

The Core Fallacy

It treats freedom of expression as though it remains the central determinant of social agency after AI severs the labor–wage–consumption circuit. Under the Discontinuity Thesis, visibility is not power, and generated speech is not productive participation. The decisive question is not merely who may speak or whether an AI has speech rights. It is who owns the models, compute, energy, distribution systems, and maintenance capacity that determine whose speech can acquire force.

Formal speech rights can survive as a legal shell after most humans lose economically necessary work. The paper studies that shell without confronting the machine capital beneath it.

Hidden Assumptions

• The liberal public sphere remains the primary arena of social power.
• Rights can be meaningfully allocated without first resolving ownership and control of AI infrastructure.
• Regulation can preserve human-centered speech domains despite coordination failure and competitive pressure.
• Users remain politically and economically decisive even as their productive role collapses.
• Raising normative questions is an adequate response to a transition governed by hard competitive mechanics.

Social Function

Classification: partial truth, prestige signaling, transition management, and ideological anesthetic.

The paper gives institutions a respectable vocabulary for discussing AI disruption without naming the harsher consequence: concentrated AI capital can preserve expression while eliminating broad human economic indispensability. It is not useless. Its analysis of visibility and machine-mediated communication describes real lag defenses. But it makes the transition discussable in legal language without making the underlying power transfer unavoidable.

The Verdict

This is legally literate reconnaissance, not a systemic diagnosis. It identifies who may control the microphone after AI arrives, but not who owns the factory that makes microphones—and decides whether humans are still needed to operate the society around them. Under DT logic, the paper can clarify regulatory friction and delay mechanisms. It cannot preserve mass productive participation or human sovereignty.