Its a matter of valence

The Tone You Trust Isn't the Record You're Voting On

Why "this feels reassuring" and "this is true" are two different tests — and Ward 13 keeps getting asked to grade on the first one.

Peter Karwacki — Ward 13 (Rideau-Rockcliffe) — Campaign Blog

There's a well-documented reason people trust a friend, a brand, or increasingly an AI chatbot: not because every claim they make has been individually verified, but because their responses are consistent, and the consistency happens to land in a direction the listener approves of. Say something reassuring often enough, in the same warm register, and the pattern itself starts to feel like evidence. Psychologists call the underlying shortcut a "consistency heuristic." It's not irrational — most of the time, in most relationships, predictability really does track reliability.

But it's also exactly the failure mode that shows up on the list of AI reasoning traps: mistaking a consistent tone for a verified outcome. A model that always answers confidently, always in the same reassuring register, will earn a user's trust regardless of whether any individual answer was actually correct. The fluency and the friendliness do the persuading. The facts, if anyone bothers to check them, do something else entirely.

Municipal politics runs on the same shortcut. A newsletter that consistently frames the ward as moving in the right direction, delivered in the same steady, positive-valence voice every week, builds exactly the kind of trust a friendship builds — approval of the pattern, not verification of the content.

What "consistent tone" actually covers for

Councillor King's weekly newsletter is professionally done and it is, week over week, remarkably consistent in register: measured, positive, forward-looking. That consistency is a communications skill. It is not a substitute for the underlying record, and in several places this year the two have pulled apart:

The consistent message	What the record shows
The Build Canada Homes federal partnership was framed as city housing progress.	Construction hasn't started — it's slated for fall 2026 to spring 2027. The city delivered 79 units in 2025 against its own 500-unit annual target.
Repeated calls for stronger LRT oversight, in consistent language, across his tenure.	He never sat on the Transit Commission or the Light Rail Subcommittee across that same tenure.
A 2022 resignation statement vouching for the Ottawa Police Services Board's competence.	Since then: a quietly closed independent workplace investigations office, Auditor General findings of inaccurate reporting to the board, and a spring 2026 misconduct crisis.
"Meaningful progress on ending homelessness."	Chronic homelessness rose from 813 to 1,279 people between 2018 and 2024. The 2026–2035 housing plan projects continued risk growth under both of its own modeled scenarios.

None of this requires assuming bad faith. It requires noticing that a consistent, approved-of tone and a verified outcome are two separate things, and that most of us — by design, not by flaw — are much better equipped to notice the first than to check the second.

Why this matters more in October than it usually does

Every election, some version of this dynamic plays out. What's different this cycle is that we're all getting daily practice at exactly this discernment problem somewhere else — in how we've learned, or are learning, to talk to AI systems. Ask a chatbot something and get a confident, friendly, consistently-toned answer, and the instinct to treat the tone as the proof is strong enough that entire literacy campaigns now exist to counter it. We're being trained, in real time, to ask a second question after the first one: not just "does this sound right," but "can I check it."

That's the same question worth bringing to a councillor's weekly update. Not "does this sound like good news," but "what does the underlying number, the committee record, or the audit say."

What I'm asking you to grade me on instead

I'm not asking Ward 13 to trust my tone. I've built this campaign's content around the opposite bet: that every claim I make should survive a records request, an Auditor General report, or an open-data pull, whether or not the conclusion is flattering to me. That's why the 311 bulk-closure analysis, the OCDSB property filing, the liabilities breakdown, and the LRT committee record are all published with their sourcing attached rather than asserted in a friendly paragraph and left there.

If a claim of mine doesn't hold up when you check it, that's a reason not to vote for me — not a reason to trust me more because I said it consistently.

The test I'd ask you to apply on October 26: not which candidate's updates make Ward 13 feel best week to week, but whose claims still hold up after you've pulled the record yourself.

Sources: City of Ottawa 2026–2035 Housing and Homelessness Plan; chronic homelessness count data (2018–2024); Auditor General reports on OPS Board reporting accuracy; Councillor King's July 24, 2026 newsletter; Build Canada Homes partnership announcement; Transit Commission and Light Rail Subcommittee membership records.