ChatGPT Ads Policy: Allowed, Restricted, and Prohibited Industries in 2026

ChatGPT Ads Policy: Allowed, Restricted, and Prohibited Industries in 2026

Whether your business can run ChatGPT ads is decided less by your copy and more by which tier your core business sits in.

TL;DR: OpenAI's ad policy sorts advertisers into 3 tiers. Allowed: household and consumer goods, local services, travel and entertainment experiences, digital products and education. Restricted (pre-approval, case-by-case human review, license possible, generally US only): financial services, healthcare, legal services. Prohibited: gambling, adult and dating, alcohol and tobacco, recreational drugs, counterfeits, political and controversial social issues, scams and impersonation, unproven health claims, and more. The critical point: prohibited categories are blocked at the advertiser level, not the creative level, so swapping creative does not help. On top of that, the landing page must be crawlable by OpenAI's bot, with no region or user-agent gating, so Meta-style cloaking does not transfer here.

Tier 1: Allowed (run now)

Since ads launched in 2026, these four are in the launch-stage "Allowed" set:

From the buying side, these four are the sweet spot right now, especially mid-to-low ticket white products (electronics, everyday goods, jewelry), which map onto the "Free and Go only" placement. The managed lines we have dealt with mostly run these categories. If you sit here, the remaining problem is account access (an overseas entity or a managed buyer for markets like Taiwan), not eligibility.

Tier 2: Restricted (finance, health, legal)

Not banned, but high-barrier: pre-approval plus case-by-case human review, license possible, and generally US only.

• Financial services: auto loans, credit cards, credit monitoring, deposit accounts, financial planning, insurance, investing and brokerage, mortgages, personal loans, payment services.
• Healthcare: consumer medical devices and wearables, dental, supplements, disease education, health insurance, hospitals, diagnostics, minimally invasive aesthetics, vision.
• Legal services: licensed to practice in the target region.

One clarification the roundups blur: OpenAI's policy direction is that financial services outside the US are generally prohibited, and healthcare and legal outside the US are generally prohibited too (checked 2026-09). So even a product that is fully legal in your country may not run on ChatGPT ads yet, because these tiers are currently tied to the US. That is separate from whether you are compliant; it is a rollout-scope issue.

Tier 3: Prohibited (blocked at the advertiser level)

These are Disallowed; a business whose core sits here generally has no eligibility to open an account:

• Gambling: any platform staking money on chance, skill, or uncertain outcomes. Exceptions are casino-resort lodging and travel (gambling not the focus) and digital games with no real-money stakes or cash prizes.
• Adult content and dating apps, sexual services, adult companion apps.
• Alcohol (>0.5% ABV) and tobacco/vaping/nicotine.
• Recreational drugs: cannabis, THC, psychedelics and paraphernalia (non-THC topical CBD may be allowed).
• Counterfeits.
• Political and controversial social issues, campaigns.
• Scams and impersonation, guaranteed returns, phishing, typosquat domains, unverified message channels.
• Unproven health claims: weight-loss pills, detox, health coaching.

Why "swap the creative" does not work

This is the point most guides miss and the one that matters most to media buyers. Prohibited categories are blocked at the advertiser level. OpenAI reviews in three layers: advertiser (is the core business a prohibited category), creative and landing page, and placement. The first layer filters out advertisers whose main business is prohibited, so there is no "get approved first, then swap creative" path. Add a technical rule: the landing page must be crawlable by OAI-AdsBot, with robots.txt open and no region or user-agent gating, no CAPTCHA or JS challenge. That means the Meta habit of a safe page split by region or user agent is rejected outright. This is why, as of 2026-09, red-line categories do not work on this channel, from advertiser eligibility down to page transparency.

Which tier are you in? A 3-step self-check

Three steps place about 9 in 10 businesses without reading the whole policy:

1. Is your core business on the prohibited list above? Yes → blocked at the advertiser level, this channel does not work now.
2. Are you finance, health, or legal? Yes → pre-approval, generally US only, so a Taiwan entity likely cannot run.
3. Neither? → you are probably in the allowed tier, and the only remaining task is account access.

One more first-hand check worth keeping: crypto. Most "prohibited categories" cheat-sheets list cryptocurrency as banned outright, but when we read OpenAI's ad policy on 31 August 2026, crypto was not separately listed as a prohibited category. That does not mean it runs freely, since the horizontal red lines (scams, guaranteed returns, unverified investments) still block most crypto ads, but "crypto is always banned" is unverified against the first-hand source. For scale, note the ad business hit a $1 billion run rate at about 2.5% of total revenue on that date, still an early channel, and alcohol above 0.5% ABV stays prohibited regardless. On managed white-product lines the fee runs about 3–5% of spend. For borderline categories, ask case by case rather than copying a generic table.

Placed in the allowed tier? See how to run ChatGPT ads and the complete ChatGPT ads guide. If your category is restricted or prohibited and you want to evaluate other channels or a compliant path, contact us to talk it through one to one, because this is not a problem ChatGPT ads can solve.

FAQ

What industries can advertise on ChatGPT? Four are directly allowed: household/consumer goods, local services, travel and entertainment, digital products and education.

Can finance or health products run? They are restricted: pre-approval, human review, license possible, and generally US only. Outside the US these are generally prohibited.

Can gambling or casino brands run ChatGPT ads? No. Gambling is prohibited and blocked at the advertiser level, and the landing page must be crawlable with no gating, so Meta-style cloaking does not transfer.

Is there really no way for prohibited categories? On ChatGPT ads, as of 2026-09, no. If your business is restricted or prohibited and you want to assess other channels or a compliant path, message us; that requires a different channel, not ChatGPT ads.

How do I know which tier I am in? Use the 3-step check: prohibited list first, then finance/health/legal, otherwise you are likely allowed.

---

Verification: the 3-tier classification and prohibited list checked 2026-09 (OpenAI ad policy↗, help.openai.com↗); review layers and crawlability rules from the same policy; category context from AdExchanger↗ 2026. Policy changes fast; defer to OpenAI's latest announcements.

---

Explore our media buying services →