AI Policy

TEXT START: This policy explains my use of “Artificial Intelligence” in my professional work.

The Dissection

This is a personal refusal policy disguised as an employability strategy. It assembles genuine harms—scraping, privacy violations, copyright abuse, environmental costs, security failures, deskilling, and unreliable output—and uses them to establish a professional boundary: the author will not knowingly generate paid work with AI.

The policy’s deeper function is identity preservation. It protects the author’s image of craftsmanship and accountability while conceding that clients, colleagues, and competitors may continue using the machinery. The contradiction is explicit: AI is rejected as a production tool but tolerated when its output arrives through other people. Responsibility is being treated as transferable by provenance, even when the author integrates the result into professional work.

The Core Fallacy

The central error is confusing present AI failures and ethical contamination with a durable economic defense against AI.

Under Discontinuity Thesis mechanics, the relevant question is not whether current models hallucinate, leak data, produce slop, or require expert supervision. The question is whether AI-assisted production can eventually deliver acceptable results at lower cost and greater speed than human-only production. If it can, abstention does not preserve employability. It creates a cost disadvantage.

The policy also mistakes the need for expertise in evaluating AI output for the survival of traditional production. Verification, security review, design judgment, and accountability may remain valuable, but they are servitor functions around AI capital—not proof that manual coding, writing, and design remain the dominant economic activity.

Evidence of subsidies, failed deployments, bubbles, and poor short-term productivity does not refute the thesis. It describes an immature and mismanaged transition. Temporary inefficiency is not a permanent moat. The steam engine did not need to be flawless for manual muscle to lose strategic importance.

Hidden Assumptions

• Clients will continue paying enough for human-only production to offset its slower throughput and higher cost.
• Current weaknesses in AI output will remain economically decisive rather than being reduced by better models, tooling, testing, and human verification.
• Professional standards are tied to the production method rather than to the quality, security, and accountability of the final result.
• Refusing AI will signal integrity and employability instead of signaling reduced capacity to compete.
• AI use can be meaningfully avoided in a sector whose infrastructure, search systems, hosting platforms, and clients are increasingly AI-mediated.
• AI-generated work can be integrated without accountability simply because the author did not initiate the generation.
• Copyright, privacy, environmental, and regulatory objections will halt adoption rather than constrain, redirect, or formalize it.
• The continuing need for expert oversight means continuing mass demand for conventional labor.
• Economic value will remain attached to producing artifacts rather than owning clients, distribution, infrastructure, verification systems, or AI capital.

Social Function

This is a partial truth performing as prestige signaling and ideological anesthetic.

Its factual criticisms are substantial, but the policy converts structural displacement into a personal purity ritual. It lets the author retain moral and professional agency while the underlying ownership structure advances unchanged. “I will not participate” is emotionally powerful but economically inert unless backed by scarce control over clients, distribution, infrastructure, or a niche where human provenance commands a durable premium.

The document also functions as a lullaby for skilled labor: it implies that maintaining standards and refusing degraded tools can preserve employability. That may work temporarily in selected markets. It does not solve the competitive pressure created when other producers become faster, cheaper, and sufficiently competent.

The Verdict

The policy is ethically coherent but strategically inadequate. It identifies real damage caused by AI while mistaking refusal for a moat. Under the Discontinuity Thesis, human-only production is a declining craft position, not a system-level defense. The author may survive temporarily through client trust, regulated work, verification, security, or a premium human-authored niche—but abstention alone is not sovereignty. It is a principled handicap attached to a shrinking labor market.