The Unspeakables¶

Handle With Care

A Warning¶

These questions are called "unspeakable" because asking them openly can end careers, destroy relationships, and mark you as a troublemaker. They are the questions that people whisper, wonder about privately, or take to their graves.

They are also often the most important questions.

This document exists because someone needs to give voice to what's unsaid. But use it carefully. Some of these questions should be asked—just perhaps not by you, and perhaps not in a meeting.

The Career Questions¶

U1. Is this project a career death trap?¶

What you're really asking: Am I being set up to take the fall for something designed to fail?

Signs this might apply: - Previous project leads have moved on quickly - Senior people are keeping visible distance - The project has political sponsors but no operational champions - Success criteria keep shifting - You were assigned rather than chosen

What you can do: - Clarify accountability in writing - Document your concerns and decisions - Build allies who will vouch for your efforts - Know your exit options - Don't be the only one holding the bag

U2. Am I the scapegoat?¶

What you're really asking: When this fails, is my role to absorb the blame?

Signs this might apply: - You're the most senior person without political protection - You're the newest person without historical excuse - You're making decisions that should be made higher up - You're signing things that should have other signatures - People are careful about what they put in writing to you

What you can do: - Ensure decisions are visibly shared - Create paper trails that show who knew what - Don't accept solo accountability for group decisions - Find out what happened to predecessors - Prepare your narrative before you need it

U3. Why am I really here?¶

What you're really asking: Was I chosen for my capabilities, or for some other reason?

Signs this might apply: - You weren't obviously the most qualified - Others seemed surprised by your appointment - Your role feels symbolic - You're not being given meaningful authority - Your input doesn't seem to affect decisions

What you can do: - Ask directly (if you have a trusted relationship) - Seek to add genuine value regardless - Build your own credibility - Don't let tokenism define you - Consider whether this is where you should be

The Honesty Questions¶

U4. Is the business case real?¶

What you're really asking: Was this designed to justify a decision already made?

Signs this might apply: - Business case was written after the decision - Assumptions are optimistic to the point of fantasy - Alternative options weren't seriously evaluated - ROI depends on perfect execution - No one believes the numbers privately

What you can do: - Stress-test assumptions explicitly - Document the real range of outcomes - Make uncertainty visible - Prepare for reality, not the spreadsheet - Don't let the fake business case become your promise

U5. Does anyone actually think this will work?¶

What you're really asking: Is everyone pretending to believe something no one believes?

Signs this might apply: - Enthusiasm is performed, not genuine - Skeptics have gone quiet rather than convinced - Eye-rolls happen outside meetings - The people closest to reality are the least optimistic - "Commitment" is mentioned frequently

What you can do: - Create safe channels for honest input - Listen to what people say off the record - Trust actions more than words - Don't contribute to the performance - Find allies who share your assessment

U6. What's the real reason for the timeline?¶

What you're really asking: Is this timeline based on reality, or politics/ego/promises?

Signs this might apply: - Timeline was set before work was understood - No one can explain why this date specifically - Technical team looks pained when timeline is discussed - Timeline matches political events (budget, election, announcement) - "Aggressive but achievable" is the phrase used

What you can do: - Ask for the basis of the timeline - Document technical reality vs. commitment - Create visibility of the gap - Prepare for what happens when timeline slips - Don't promise what you can't deliver

The Darker Questions¶

U7. Is someone lying to me?¶

What you're really asking: Is there deliberate deception happening, not just optimism?

Signs this might apply: - Stories don't add up - Information changes depending on who's asking - Documents differ from verbal statements - Key information is withheld - People avoid putting things in writing

What you can do: - Seek independent verification - Create written records of what you're told - Cross-reference information sources - Trust your instincts about inconsistency - Consider the implications if it's true

U8. Is something illegal happening?¶

What you're really asking: Am I at risk of being part of something that crosses legal lines?

Signs this might apply: - Processes are being bypassed - Documentation is deliberately vague - People are uncomfortable but compliant - Normal oversight is circumvented - There's pressure to "just do it"

What you can do: - Know the relevant laws and policies - Document your concerns - Seek legal advice privately - Understand whistleblower protections - Know the difference between uncomfortable and unlawful

U9. Is someone benefiting inappropriately?¶

What you're really asking: Is there corruption or conflict of interest I'm not seeing?

Signs this might apply: - Vendor relationships seem unusually close - Procurement decisions don't make sense on merit - Someone is pushing hard for a specific outcome - Decision-makers have undisclosed interests - Things that should be competitive aren't

What you can do: - Observe patterns without accusing - Understand conflict of interest policies - Ensure your own integrity is unimpeachable - Know how to raise concerns safely - Don't participate in what you can't defend

The Existential Questions¶

U10. Is this AI actually solving the right problem?¶

What you're really asking: Have we defined the problem correctly, or are we solving what we can measure?

Signs this might apply: - The problem statement emerged from the solution - AI is doing what's measurable, not what matters - Success metrics are disconnected from citizen outcomes - The real problems are social/political, not technical - No one has asked the affected people what they need

What you can do: - Return to first principles - Ask what problem affected people would prioritize - Distinguish problem from symptom - Be honest about what AI can and can't address - Don't let the solution define the problem

U11. Should this AI exist at all?¶

What you're really asking: Are we building something that shouldn't be built, regardless of how well we build it?

Signs this might apply: - Use case is fundamentally questionable - Ethical concerns are dismissed rather than addressed - Similar systems have caused harm elsewhere - The power imbalance created is concerning - Your conscience is uneasy

What you can do: - Take ethics seriously, not as a checkbox - Learn from what similar systems have done - Consider the precedent being set - Ask whether regulation would allow this - Listen to your conscience

U12. Am I part of something harmful?¶

What you're really asking: Even if this succeeds by its own measures, will it do more harm than good?

Signs this might apply: - Benefits accrue to the powerful, costs to the vulnerable - Efficiency is valued more than humanity - The system creates more control than assistance - Citizens are subjects, not partners - You'd be uncomfortable if your family was affected

What you can do: - Assess full impact, not just claimed benefits - Consider who bears costs vs. who receives benefits - Apply the "would I want this for my family" test - Don't rationalize what troubles you - Consider whether your participation is essential

The Political Questions¶

U13. What does the Minister actually want?¶

What you're really asking: What's the real objective, beyond the stated one?

Signs this might apply: - Ministerial interest is intense but vague - Political timing seems to matter more than operational reality - The project is more about announcement than outcome - Success is defined by political optics - Technical decisions defer to political considerations

What you can do: - Understand the political context - Protect operational integrity where possible - Don't let political objectives override good practice - Document when political pressure affects decisions - Know the difference between serving the Minister and serving the public

U14. Who in this room is being political?¶

What you're really asking: Who is saying what they believe vs. what serves their interests?

Signs this might apply: - Positions correlate with interests - Arguments seem performative - The same people always agree with power - Genuine concerns are coded in careful language - What's said in meetings differs from what's said afterward

What you can do: - Read the room - Seek private conversations for real views - Understand incentives shaping behavior - Build relationships that enable honesty - Be aware of your own political positioning

U15. Is this about helping citizens or about something else entirely?¶

What you're really asking: Is citizen benefit the actual purpose, or the justification for another purpose?

Signs this might apply: - Citizen impact is mentioned but not measured - The project serves institutional interests more obviously - "Efficiency" and "savings" dominate "service" and "outcomes" - Citizens weren't consulted meaningfully - Success is defined by the system, not the citizen

What you can do: - Re-center citizen outcomes - Measure what matters to affected people - Speak for citizens when they're not in the room - Challenge efficiency-first framing - Ask whether this would exist if citizens had designed it

How to Handle Unspeakable Questions¶

Thinking Them¶

These questions should be thought about privately, honestly, regularly. Your conscience is a legitimate source of information. If something feels wrong, examine why.

Speaking Them¶

Not all questions can be spoken in all contexts. Consider: - Who can you ask safely? Trusted allies, mentors, people with nothing to lose - How can you ask? Indirect framings, hypotheticals, "what would a skeptic say?" - Where can you ask? Private conversations, anonymous channels, external advice - Whether to ask at all? Some questions are better answered by observation than direct inquiry

Acting on Them¶

If unspeakable questions reveal unspeakable truths: - Document what you know and how you know it - Seek advice from trusted sources - Understand your options before you need them - Protect yourself with records and relationships - Know when to act and when to wait - Know when to leave if you cannot stay with integrity

The Final Unspeakable Question¶

U16. Should I be doing this?¶

What you're really asking: Given everything I know and suspect, is my participation ethical?

There is no framework for this question. There is only your judgment, your conscience, and your willingness to live with the consequences of your choices.

Some people stay and fight from within. Some leave and speak from outside. Some stay and comply. Some leave in silence.

There is no universal right answer. There is only your answer.

But the question deserves to be asked. Because participating in harmful systems while telling yourself you have no choice is itself a choice.

"The questions we cannot ask are often the questions that matter most. The unspeakable questions exist whether we speak them or not. The difference is whether we face them with our eyes open."

A Note on Safety¶

If you are genuinely concerned about illegality, corruption, or serious harm:

• Understand whistleblower protections in your jurisdiction
• Seek legal advice before acting
• Document carefully and securely
• Consider external channels if internal ones are compromised
• Protect yourself—the system often punishes truth-tellers

These are not theoretical risks. People have lost careers, health, and more for speaking truth. The fact that it's the right thing to do doesn't make it safe.

Know the risks. Then decide.