New York Governor 2026: the ads, and who's paying for them
Kathy Hochul (Democrat) and Bruce Blakeman (Republican) are the candidates for governor of New York in the November 3, 2026 general election. Kathy Hochul is the incumbent. Semblen has analyzed 76 ads in this race; 44 used at least one manipulation technique, most often loaded language and name calling.
Kathy Hochul (DEM, incumbent), Bruce Blakeman (REP).
On the ballot as listed by Ballotpedia. 76 ads analyzed, 44 with at least one technique in the copy.
Newest first. The paid-for-by line is the advertiser's own filing. Updated 2026-10-08.
42 of 74 (57%) flagged · Loaded Language (31), Name Calling (21), Appeal to Fear (17)
YouTube
2 of 2 flagged · Bandwagon Appeal (2), Appeal to Fear (1), Appeal to Prejudice (1)
Reported Meta spend
$333,300–432,067 on all ads, $196,300–251,366 on flagged ads · Meta reports ranges; these are the sums of their ends
Example
“High Taxes, Rising Utility Rates” — Loaded Language, paid for by Blakeman for New York
· the flagged ad with the highest reported spend that shows the most common technique
Embed this race on your site
A small box with these findings and the newest ads, updated as new ads are analyzed. Free for newsrooms, blogs and campaigns alike; it links back here.
Follow Kathy Hochul vs. Bruce Blakeman
An email when new ads come in from either side or the outside groups: the techniques, the words doing the work, and who paid. And each debate's analysis the morning after. One-click unsubscribe.
Each side is the campaign plus the groups backing it. How a group is placed is explained below the table.
Side
Ads
With at least one technique
Most common
Hochul (D)
18
7 of 18 (39%)
Loaded Language (6), Name Calling (5), Appeal to Fear (2)
All from the campaign; no ads here from groups placed on this side.
Blakeman (R)
33
20 of 33 (61%)
Loaded Language (12), Glittering Generalities (10), Appeal to Fear (7)
All from the campaign; no ads here from groups placed on this side.
Not placed on a side (15 sponsors)
25
17 of 25 (68%)
Loaded Language (13), Name Calling (11), Appeal to Fear (9)
Read this before comparing the sides. Ads that attack an opponent use these techniques far more often than ads that introduce a candidate, whoever runs them, so the side running more attack ads shows a higher share. A technique is about wording, not about whether an ad is true. These are online ads only (Meta and YouTube), not television.
How a group is placed: by its own filings with the Federal Election Commission. A group is with a candidate when it reported spending to support that candidate, or to oppose the other one in the general election. The parties' own committees are with their nominee. Spending against a candidate in a primary places a group nowhere, and neither does our reading of its ads: a group with no such filing is listed as not placed. The sponsor table gives the reason for each.
Decide what counts
Not everyone would call every technique manipulation. Untick the ones you'd call ordinary persuasion and the count updates.
58% of the 76 ads here use at least one of the techniques ticked below (44 ads).
Leaving out the softer ones (loaded language, glittering generalities, slogans, exaggeration, repetition): 49% of ads.
The words doing the work
The words that show up most in phrases our model flagged as loaded language or name calling in this race's ads, by number of ads.
billin 5 ads· e.g. “skyrocketing utility bills”, “skyrocketing electric bills”
MAGAin 5 ads· e.g. “MAGA extremism”, “New York MAGA extremism”
policyin 4 ads· e.g. “pro-criminal policies”, “destructive anti-energy policies”
criminalin 3 ads· e.g. “Hochul's New York: were criminals roam free”, “criminals roam free”
hochul'in 3 ads· e.g. “Hochul's New York: were criminals roam free”, “Kathy Hochul's fraud and abuse”
pro-criminalin 3 ads· e.g. “pro-criminal policies”, “most pro-criminal governor in America”
worst governorin 3 ads· e.g. “WORST Governor in America”, “the worst Governor in America”
Semblen is free to read and has no ads. If this page was useful, you can support it.
Money in this race
Candidates for governor report their fundraising to New York's state campaign-finance office, not the FEC. This page doesn't cover those reports yet, so it shows no money figures.
The ads
76 ads, newest first. Video transcripts are shown as the model read them, without caption cues like [music].
Showing the 1 ad paid for by Friends of Matt Slater. Show all 76 ads
Paid for by Friends of Matt Slaterfrom 2026-09-24spend $600–69915,000–19,999 impressions
Kathy Hochul wants to turn New York into a sanctuary state, and it's our families who pay the price. Matt Slater is leading the fight to end sanctuary policies and put New Yorkers first. Add your name to the petition and stand with Matt.
Appeal to Fear: The sentence implies that sanctuary policies directly harm families, creating anxiety about a threat to motivate action — for example: 'our families who pay the price'
Scapegoating: Sanctuary policies (and by extension immigrants) are blamed for harm to families, attributing complex social issues to a single policy/group — for example: 'our families who pay the price'
Appeal to Prejudice: The phrase 'our families' invokes in-group identity to contrast law-abiding residents against those protected by sanctuary policies — for example: 'our families'
Glittering Generalities: The phrase 'put New Yorkers first' uses vague, emotionally positive language that sounds noble but carries no specific policy meaning — for example: 'put New Yorkers first'
Bandwagon Appeal: The call to 'add your name to the petition and stand with Matt' pressures conformity by implying others are already joining a movement — for example: 'Add your name to the petition and stand with Matt'
False Urgency: The imperative 'Add your name' combined with the framing of an ongoing 'fight' creates pressure to act immediately without deliberation — for example: 'Add your name to the petition'
Appeal to Fear
Scapegoating
Appeal to Prejudice
Glittering Generalities
Bandwagon Appeal
False Urgency
Looks up the ad's factual claims in fact-checkers, primary sources and wire reporting. Takes up to a minute.
An email when new ads come in from either side or the outside groups: the techniques, the words doing the work, and who paid. And each debate's analysis the morning after. One-click unsubscribe.
An ad is on this page because its own "Paid for by" line is a campaign in this race; or it ran mostly in NY and names a nominee for governor — unless its sponsor is a campaign for some other office (a U.S. Senate or House campaign included), or it also names NY's Senate nominees, in which case it is on the Senate race's page. Ads from outside groups appear alongside the campaigns' own.
The candidates are the ones certified for the November ballot, from the source linked at the top. Governors are state offices, so there are no FEC filings behind this page: no money figures, and outside groups are placed on a side only when they are a campaign or a party committee.
Campaigns run one text as many separate ads. Identical text from the same sponsor is one card here, marked with its number of variants.
A technique is a specific manipulative move in the wording, found by our model and explained in one line. It is not a claim that the ad is false.
Claims are checked when a reader asks: press "Check this ad's claims" on a card. The check runs once and its result is shown to everyone afterwards. The ten ads with the most money behind them in each race are checked ahead of time.
Claim verdicts say what the cited sources say. A verdict that an ad is wrong or misleading is shown only when it cites an established fact-checker or an institutional record — Congress, a federal agency, a court — and its confidence is not low; a weaker one is not shown at all. Each verdict shows what it rests on and how confident it is. If we got one wrong, press Disagree: a verdict that readers dispute and no one defends comes down.
Video ads come from the campaigns' own YouTube channels; the transcript is analyzed. Meta ads are analyzed on their text; their video and images are not, yet.
Get notified as this improves — new detection models, the browser
extension, and new features. Occasional email, no spam.
Semblen LLC · 5900 Balcones Drive STE 100, Austin, TX 78731 · unsubscribe any time