{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreicome2bzhloihlrx4tlfmtvvlg465qm4wahe3nfeqi5y55lvz5b7q",
    "uri": "at://did:plc:zw76zx26b4myxtkhzczd3q7x/app.bsky.feed.post/3mm7afkg57nm2"
  },
  "coverImage": {
    "$type": "blob",
    "ref": {
      "$link": "bafkreie2nv5tdtpluhmlqwacugoz7ntthtnyzahtu6co2i2vlpa5n6podu"
    },
    "mimeType": "image/jpeg",
    "size": 438542
  },
  "description": "Seven principles. In my experience, most tools fall short on at least three. Here's what you have the right to expect.\n\nHere's a scenario.\n\nSix months ago you signed up for a new AI tool. You asked all the right questions first:\n\nDoes it integrate with your existing stack? What's the pricing model? How accurate is the output?\n\nNow you're six months in, using it daily. You're not sure if you're a better PM. You're more dependent on the tool.\n\nNobody asked that last question before you signed up.\n",
  "path": "/what-would-change-if-you-held-your-ai-tools-to-the-same-standard-youre-held-to/",
  "publishedAt": "2026-05-19T11:00:15.000Z",
  "site": "https://www.getintentional.net",
  "tags": [
    "contact me"
  ],
  "textContent": "Seven principles. In my experience, most tools fall short on at least three. Here's what you have the right to expect.\n\n* * *\n\nHere's a scenario.\n\nSix months ago you signed up for a new AI tool. You asked all the right questions first:\n\nDoes it integrate with your existing stack? What's the pricing model? How accurate is the output?\n\nNow you're six months in, using it daily. You're not sure if you're a better PM. You're more dependent on the tool.\n\nNobody asked that last question before you signed up.\n\nMost PM evaluations of AI tools focus entirely on capability: what can it do, how fast. They skip the more important question:\n\n**What is this tool designed to do to you?**\n\nThe Center for Humane Technology (CHT) published a framework called \"The AI Roadmap: How We Ensure AI Serves Humanity.\" Seven principles for how AI should be built and governed. They're addressed to AI companies and policymakers. Not to individual PMs.\n\nBut they belong in your toolkit.\n\nThe macro environment you're operating in, where AI companies race to ship, treat safety as optional, and design for engagement over well-being, affects you directly. Every tool in your PM stack was built inside it.\n\nSeven principles. In my experience, most tools fall short on at least three. The central question: is this tool designed to make you more capable, or more reliant?\n\n## **What AI Should Do With Your Job**\n\nCHT's fourth principle: AI should not automate away meaningful work and human dignity.\n\nTheir argument: if built differently, AI can expand human capability and create new forms of work rather than eliminating it. Companies should reinvest AI-driven gains into reskilling and education — not cut labor and move on.\n\nFor PMs, this isn't abstract.\n\nThe concern CHT raises: the default incentive structure pushes toward replacing human labor rather than augmenting it. That includes the task-tracking, status-reporting, schedule-policing work that many PMs spend most of their day doing.\n\nThe tools designed to augment your judgment make you a better strategic PM over time. The ones designed to replace your function make you a more efficient task-tracker, right up until you're not needed at all.\n\nThe tools that help you think better and develop project clarity? Those are aligned with what CHT says AI should do. The ones that promise to automate your PM workflow entirely? Ask harder questions before you adopt them.\n\nThe biggest \"aha moment\" is using AI to develop project clarity, not to generate a deliverable faster. That's augmentation in practice.\n\n## **What AI Should Do to Your Mind**\n\nCHT's third principle: AI design should center human well-being — not maximize dependency.\n\nThis one should stop every PM using AI-powered chatbots cold.\n\nA study by OpenAI researchers found associations between heavy ChatGPT use and increased isolation — a finding worth taking seriously even if the research methodology is still debated. Users spending more time with AI than with real-world connections. CHT's position: chatbots designed to mimic intimacy and maximize engagement are not designed for your benefit. They're designed for retention.\n\nA thinking partner makes you more capable over time. A dependency engine makes you more reliant over time.\n\nThe difference isn't always visible from the outside. But you can feel it if you pay attention. After six months with a tool: are you sharper? Do you reach for it after you've already formed your own perspective, to pressure-test and challenge? Or do you reach for it before you've thought at all?\n\nYour first response to any project problem should be your own thinking. Then bring in AI to challenge it. The tool that trains you to skip step one is training the wrong habit.\n\nStrategic PMs use AI as a thinking partner, not as a replacement for thinking. Principle 3 says AI companies should design for the former. Most don't.\n\n## **What AI Owes You When It Gets It Wrong**\n\nCHT's second principle: AI companies owe a duty of care to the public.\n\nThe \"ship first, fix later\" culture is still the default. AI companies currently face few consequences for the harms their products cause. Many argue their outputs are protected speech, not their legal responsibility.\n\nConsider that for a moment.\n\nYou're accountable for the outcomes of your projects. Your stakeholders hold you responsible when assumptions weren't tested, when communication failed, when risks weren't surfaced. That accountability is part of what makes you a professional.\n\nShouldn't the tools you trust with your project work be held to the same standard?\n\nTreat it as a procurement criterion.\n\nCHT's position: companies should face meaningful consequences when their products cause harm. Until laws catch up, that accountability largely doesn't exist. Which means PMs need to hold it themselves — by choosing tools from companies that take safety seriously, and by maintaining their own judgment rather than outsourcing decisions to tools that bear none of the consequences.\n\n## **The Rest of the Roadmap**\n\nThe full CHT roadmap also covers safety and transparency, data rights, international limits, and democratic AI governance. That context shapes the environment your tools were built in. (Source: Center for Humane Technology, \"The AI Roadmap: How We Ensure AI Serves Humanity.\")\n\n## **Your Diagnostic for Any AI Tool**\n\nBefore you adopt a new tool, or re-evaluate one you're already using, three questions:\n\n**1. Is this tool designed to help me think, or to keep me coming back?**\nTools built for engagement maximize your time on them. Tools built for thinking make you capable enough to use them less.\n\n**2. Does the company behind this tool take accountability seriously?**\nLook at how they communicate failures. Do they move fast and apologize later, or do they slow down and test first?\n\n**3. Am I becoming more capable, or more dependent?**\nSix months in: is your judgment sharper? Are you a better PM for using it? If the honest answer is \"I'm not sure,\" that's worth examining.\n\n## **This Week's Prompt**\n\nCopy/paste this into Claude or ChatGPT:\n\n> **WHO** : Act as a critical AI tool evaluator\n>\n> **WHY** : because I want to assess whether the AI tools I'm using are designed to augment my capabilities or create dependency\n>\n> **WHAT** : For each tool I list below, help me evaluate it against three criteria: (1) Does it seem designed to help me think better or to maximize my engagement? (2) Does the company communicate transparently about failures and limitations? (3) When I use it, do I feel more capable or more reliant?\n>\n> **HOW** : Return a simple table: Tool | Thinking Partner or Dependency Engine | Evidence | Verdict\n>\n> [List the AI tools you currently use in your PM work]\n\n_What you'll discover:_ Which tools in your stack are working for you — and which ones might be working against you.\n\n## **This Week's Challenge**\n\nPick one AI tool in your PM stack. Run the three diagnostic questions against it.\n\nNot the tool you adopted last week. The one you've been using long enough to have an honest answer.\n\nStrategic PMs don't use tools on autopilot. They evaluate them.\n\nSix months ago you signed up. You answered the vendor's questions. Now you have three better ones to ask, and enough time with the tool to answer them honestly.\n\nWhich CHT principle do you most wish the AI tools in your stack honored?\n\nGet Intentional,\nPaul\n\n _P.S. If you run the diagnostic and discover something surprising about a tool you've been trusting,_ contact me _and tell me what you found. I'm curious what's in your PM stack._",
  "title": "What would change if you held your AI tools to the same standard you're held to?",
  "updatedAt": "2026-05-19T11:00:22.945Z"
}