AI Usage Policy

Author: David Bushell — Web Design & Front-end Development (UK) Site: dbushell.com | valleyfold.co.uk

This document outlines my professional stance and operational guidelines regarding the integration of "Artificial Intelligence" into my workflow.

Definition: For the purposes of this policy, "AI" specifically refers to the current trends in Large Language Models (LLMs) and Generative AI.

Timeline:

• Initial Publication: 2025-06-18
• Latest Revision: 2026-05-05

---

⚡ Quick Overview

• Current Status: I do not utilize AI for any professional engagements.

Executive Summary

I commit to not knowingly employing AI to generate any professional deliverables, including but not limited to:

• Source code
• Visual design
• Client or professional correspondence

I strive to minimize my interaction with AI tools whenever possible while producing paid work. However, given the aggressive and pervasive nature of AI deployment, total avoidance is impractical.

Collaboration Terms:

1. I will continue to partner with clients and peers who choose to use AI.
2. If I must integrate AI-generated content provided by a third party, it is under the strict condition that I am not held accountable for the integrity or quality of that specific combined output.

---

🧠 Rationale

This policy is a pragmatic approach to maintaining my personal professional integrity while remaining viable in a volatile economic climate. My objections are categorized as follows:

1. Morals and Ethics

The AI industry operates through methods that I find unacceptable:

• Infrastructure Abuse: AI crawlers utilize deceptive tactics (e.g., residential proxies and fake user agents) to scrape data, ignoring the industry-standard robots.txt file.
• Web Degradation: This has led to massive server strain and the proliferation of intrusive CAPTCHAs to fight DDoS-like traffic.
• Societal Harm:
  • "AI slop" is flooding social media, fueling scams and misinformation.
  • AI-driven search results often prioritize hallucinated summaries over authentic sources.
  • The technology is used to create sexual abuse material and has, in some cases, encouraged vulnerable individuals toward suicide.
  • The forced integration of chatbots into software ignores the potential for inducing psychosis.
• Human & Environmental Cost: Data centers cause significant air, noise, and water pollution. Furthermore, the "cleaning" of training data has subjected low-paid contractors to severe psychological trauma.

The Copyright Paradox: AI firms have built empires on datasets harvested without consent. While some claim "fair use," many large media companies are signing lucrative licensing deals.

$\text{If Fair Use} = \text{True} \implies \text{Why pay for licenses?}$

If training is truly fair use, then copyright is effectively being "laundered" through AI to erase legal protections.

Conclusion: I cannot use AI in good conscience while simultaneously valuing my own creative rights and the health of the web platform.

2. Privacy and Security

The deployment of AI has introduced critical vulnerabilities:

• Data Leaks: Confidential information is often absorbed into models with no clear path for deletion or opt-out.
• Deceptive Patterns: Pro-consumer legal requirements are being bypassed to prioritize AI revenue.
• Technical Threats:
  • Prompt injection remains a persistent threat with no foolproof fix.
  • AI-generated code frequently introduces CVEs (Common Vulnerabilities and Exposures) rated as High or Critical.

Conclusion: Utilizing AI risks breaching client trust and violating UK/EU data protection laws, regardless of signed NDAs.

3. Quality and Standards

From a technical perspective, AI output is insufficient:

• Subpar Code: AI-generated code fails to meet my professional benchmarks and the high standards my clients expect.
• Bias: System prompts often push specific technologies that are inappropriate for a given project.
• Design Flaws: Using Generative AI for visuals is essentially like playing a $\text{slot machine}$—it is random and unreliable.

Conclusion: To deliver AI-generated code or design would be an act of professionalism professional negligence.

4. Employment, Education, and Economy

The broader impact on the industry is devastating:

The "Cognitive Debt" Equation: Long-term reliance on LLMs leads to a decline in critical thinking: $\text{Cognitive Debt} \approx \sum (\text{AI Reliance}) - \text{Active Problem Solving}$

Furthermore, the AI boom has artificially inflated the cost of essential hardware, such as GPUs and memory.

Conclusion: AI markets itself as a tool that removes the need for skill, yet it ignores the expert knowledge required to verify and fix its errors. This accelerates the "deskilling" of web development and blocks new talent from entering the field.

---

👤 About David Bushell

Based near Manchester, UK, I specialize in the design and construction of web experiences. I founded Valley Fold to transition from freelance work toward a mission of helping real people enjoy a better, more human web.

Connect with me:

• Follow: Mastodon | Bluesky
• Stay Updated:
  • Subscribe to Blog RSS
  • Subscribe to Notes RSS
  • Combined Feed

Search: [____________________] [Searching...]