The question arrived in our inbox four times last week. Same query, four different readers, two of them forwarding the same Reddit screenshot of a Google Ads disapproval notice supposedly tied to an Ohio prediction market campaign. The screenshot was the only evidence offered.
We need to say something before we go further. Our investigative desk works from primary documents — enforcement notices on the UKGC public register, MGA sanction lists, AGCO bulletins, NJDGE adjudications, annual reports filed with regulators that have subpoena power. Google's internal ad-policy decisions are not in any of those registers. They are not filed. They are not, in any meaningful sense, on the public record the way a £17m regulatory settlement is on the public record. So when the question lands — *did Google ban prediction market ads in Ohio?* — the honest answer from this desk is that we cannot verify the specific Ohio ad-block decision from a primary source in our dataset.
That gap is the piece. The gap between an unsignposted platform policy and a published regulatory action is the entire reason prediction markets sit where they sit in the US legal landscape. And it is the same gap that licensed operators have worked profitably for two decades.
The Question Arrives Without a Primary Document Attached
Here is what we do have. We have a reader pattern. Four queries in eight days, three from Ohio IP ranges, one from Kentucky. All four reference the same Polymarket-Kalshi-aligned ad-policy issue. None of them sent a Google policy URL — because Google does not publish geographically-scoped enforcement of its gambling ads policy the way a regulator publishes a fine. There is no Ohio-specific Google Ads transparency document we can quote. There is a global gambling ads policy page that says certain content is restricted by jurisdiction, and there is a long history of platform-level enforcement that adjusts quietly in response to state attorneys general making noise.
The CFTC's jurisdictional turf war with state regulators over event-contract markets is real, well-documented elsewhere, and not in our dataset. We will not invent it for you. What we can do is tell you what the absence of a primary document means in this kind of question.
It means the enforcement is opaque by design. A platform ad-block does not create a public record. It does not produce a settlement statement. It does not enter a register that a researcher can pull six months later to verify what happened. A regulator-led enforcement action — the kind we track — does the opposite. When the Gambling Commission settled with Entain in 2022 for £17m, the public regulatory settlement statement named the failure pattern in writing: insufficient customer interactions with high-risk players, inadequate identification of problem gambling indicators, AML controls insufficient for unusual deposit patterns. A reader can quote that document back to the operator three years later. A reader cannot do the same with a Google ad disapproval notice.
The UKGC public register loads slowly on Friday afternoons. It still loads.
The asymmetry matters because it shapes what readers can demand. From a regulator, you can demand the underlying enforcement notice. From a platform, you can demand a help-center article that says "this content type may be restricted in certain regions." Those are not the same instrument, and treating them as the same instrument is what gets readers into trouble when they try to build a position on what is actually happening to prediction markets in Ohio.
What Enforcement Looks Like When a Regulator Actually Acts
Let us put a real document on the table to show what a primary record looks like in this industry, so the contrast with the Google question lands properly.
In March 2023 the UK Gambling Commission fined a Flutter UK licensee £1.17m for social responsibility and anti-money laundering failures at Sky Betting and Gaming. The notice names the operator, names the failure category, sets the financial penalty, and enters the public register. A reader in Toronto can pull it. A compliance officer in Stockholm can pull it. We can pull it three years later and quote the date — 2 March 2023 — and the amount, and the scope, without paraphrase. That is what enforcement looks like when the system works the way it is meant to work.
Eight months later, the same parent group's competitor cleared a different threshold. Entain announced a Deferred Prosecution Agreement with the UK CPS carrying a £585m settlement tied to the former Turkey-facing business of Headlong Limited, a subsidiary the group had sold in 2017. Six years between the disposal and the settlement. The DPA document is dated, public, and quotable.
Now hold that DPA against Entain's 2024 annual report. The same operator reports £4,833m in revenue and 28 million active customers, with 88% of group revenue from regulated markets. Both documents are on the public record. Both are operative. Both describe the same company. One says "regulated markets are where we are" and the other says "we paid £585m to settle a criminal investigation into a market we exited seven years ago." The gap between them is not a contradiction. It is the time signature of how enforcement actually operates against a tier-1 licensed operator: years late, quantified, on a public register, with a precise scope statement.
That is the standard a reader should hold the prediction market question to. Where is the regulator who has put their name on a published enforcement document? Where is the dated notice? Where is the scope statement?
In the prediction market case in Ohio, as far as our register goes, there is none. There may be one tomorrow. There is not one today that we can quote.
Our enforcement-tracking spreadsheet has 47 columns. None of them are "Google ad policy."
The Prediction Market Gap Sits in the Same Crack Licensed Operators Have Always Worked
Here is the part that connects the question back to the industry we actually cover. The reason prediction markets are in this ad-policy limbo at all — the reason a platform like Google is making the call instead of a regulator — is that the US gambling enforcement architecture is jurisdictionally fragmented in a way that creates exactly the same exploitable seams licensed operators have spent twenty years mapping.
FanDuel operates in 22 US states under state-by-state licensing, with 28.5% market share in New Jersey under NJDGE supervision. DraftKings operates in 27 states. Both are publicly traded. Both report quarterly. Both sit inside enforcement jurisdictions where a state regulator can write a notice that ends up on a public register. Their entire compliance posture is built around that fact — segregated player funds, RNG certifications from Gaming Laboratories International, eCOGRA dispute mediation, license tier disclosures in 10-K footnotes. They live inside the regulated perimeter because the regulated perimeter is what they sold their public shareholders.
Prediction markets — Kalshi, Polymarket, the CFTC-supervised event-contract category — sit outside that perimeter by design. They are not state-licensed gaming operators. They are CFTC-supervised derivative venues, which means the state attorneys general who police gambling ads do not have direct jurisdiction over the underlying product, only over the marketing of it inside their state. That jurisdictional gap is what produces the ad-policy whack-a-mole the reader is asking about. A platform decides to block ads in a state because a state AG sent a letter. The state AG sent the letter because the federal regulator did not assert exclusive jurisdiction clearly enough to preempt the state's ability to act. None of that produces a public enforcement document on a register we can pull.
Look at how a clean jurisdiction handles this in contrast. AGCO Ontario supervises 49 licensed iGaming operators inside a single regulatory perimeter with a single registration system, a single advertising standard, and a single enforcement body. An operator that crosses the advertising line gets named in an AGCO bulletin. The bulletin is dated, scoped, public. Compare that to the US where a prediction market lives across three regulators (CFTC, state AGs, the platform itself acting as quasi-regulator) and gets enforced against by the one with the least documentation discipline. That is not a bug. It is the structural choice the US made when it built a gambling enforcement system one state at a time.
Four readers, four different Reddit threads, same screenshot. None of them with a regulator's name on it.
The reader who emails us asking about the Ohio ad block is asking the right question but at the wrong layer. The interesting question is not whether Google blocked the ads. The interesting question is why Google is the entity making that decision in the first place, instead of the Ohio Casino Control Commission filing a public order. And the answer to that is the same answer that explains why the US iGaming map is a patchwork of state regulators with material enforcement-disclosure differences — and why an operator with a Curaçao license can run ads in some US states that an AGCO-licensed operator would never be permitted to run in Ontario. The architecture decides the enforcement. The architecture in this case is not built for clean documentation.
---
This started as a one-line answer to a reader email about an Ohio ad-policy decision and turned into a piece about why ad-policy enforcement sits at the weakest, least-documented layer of the entire gambling regulatory stack. We did not get to verify the Ohio claim from a primary document. The fact that we could not is the most useful thing the question revealed, because it is the same gap that lets a £17m UKGC settlement take three years to materialize against an operator that the public register knew was running hot the whole time. The reader who wanted a yes or no got a different answer. We think it is the more useful one.
FAQ
Did Google actually ban prediction market ads in Ohio?
Our investigative desk could not verify this specific claim from a primary regulatory document in our dataset. Google's internal ad-policy enforcement does not appear on a public register the way a UKGC fine or an AGCO bulletin does. Readers forwarding screenshots of disapproval notices are reporting a real platform-level action, but the absence of a regulator-issued document means the scope, duration, and legal basis are not externally verifiable. Treat platform ad-policy decisions and regulator-issued enforcement orders as different instruments with different evidentiary weight.
Why does the prediction market sit in this regulatory gray zone at all?
Prediction markets operate as CFTC-supervised event contracts rather than state-licensed gambling products. That federal classification creates jurisdictional friction with state attorneys general who police gambling advertising at the state level. The friction is not resolved by either side cleanly, so platform intermediaries — payment processors, ad networks, app stores — end up making operational decisions that look like regulation but produce no public enforcement record. This is the structural source of the documentation gap readers run into when they try to verify what happened.
How does this compare to how licensed sportsbooks are regulated in the US?
Licensed operators like FanDuel and DraftKings work inside a state-by-state regulatory architecture. FanDuel operates in 22 states; DraftKings in 27. Each state issues its own license, sets its own advertising rules, and publishes enforcement actions through its state gaming commission. New Jersey's NJDGE, for example, publishes adjudications under its Division of Gaming Enforcement. That is a fundamentally different enforcement posture than the prediction market category, which sits federal-supervised but state-marketing-restricted, with no single register collecting the actions taken against it.
What does a real published enforcement action look like in this industry?
A useful contemporary example: in August 2022 the UK Gambling Commission published a £17m regulatory settlement statement against Entain covering social responsibility and anti-money laundering failings at Ladbrokes and Coral. The statement names the failure types — insufficient interactions with high-risk players, inadequate problem-gambling identification, AML controls insufficient for unusual deposit patterns. The document is dated, scoped, on the public register, and quotable by a researcher years later. That is the documentation standard a reader should expect when they ask whether a regulatory action occurred.
Why does the publication say it cannot verify some claims even when readers have screenshots?
A screenshot of a platform notification is evidence that something happened to that user. It is not a primary regulatory document. Our desk works from filings — annual reports, license registers, enforcement notices, deferred prosecution agreements — that carry the weight of being publicly filed under a regulator's signature and verifiable years later by any researcher. We will tell you when a question falls outside the perimeter of what those documents cover, because the alternative is to invent the documentation, and the entire premise of being taken seriously rests on not doing that.
Where can a reader actually verify gambling enforcement actions themselves?
For UK-licensed operators, the Gambling Commission's public register lists every active licensee and every published regulatory settlement; the register currently shows 268 licensed online operators. For Ontario, AGCO publishes operator bulletins and oversees 49 licensed iGaming operators. For New Jersey, the NJDGE publishes adjudications. For listed operators, the annual report — Entain's 2024 filing, Flutter's results centre — is the source of record for material regulatory matters. These are the four highest-signal primary sources for English-language gambling enforcement research.
What should a prediction market user do with this kind of opaque enforcement environment?
Treat platform-level ad blocks, payment processor restrictions, and app store removals as leading indicators rather than legal determinations. They typically precede formal regulatory action by months or sometimes years, and they reflect the platform's risk assessment more than a finding of unlawfulness. A user wanting to understand their actual legal exposure in a given state should look to the state attorney general's published statements and the relevant federal supervisor's no-action guidance — not to whether a Google ad displayed in their feed last Tuesday.