ART Act 2021 — IVF and ART ChatGPT Ads compliance
The ART (Regulation) Act 2021 was written before conversational ad units existed, but its restrictions on IVF and fertility advertising apply to a ChatGPT Ad exactly as they apply to a hoarding or a Google search ad. ICG runs IVF and ART clients' ChatGPT Ads through a compliance screen tied to the Act's banned-language list, mandatory disclaimers, and National Registry linkage — before a single variant serves.
What the rule actually says
The Assisted Reproductive Technology (Regulation) Act, 2021, notified in January 2022, is the primary statute governing IVF, gamete donation, and ART bank operations in India. Its advertising provisions sit mainly under Sections 32 through 35, and read together they establish a fairly narrow set of things an ART clinic or bank is permitted to say publicly about its own services, regardless of the medium carrying the message. The Act does not carve out digital advertising, paid search, social media, or conversational AI as a separate category — it speaks to "advertisement" in the broadest sense, and the intent of the drafting, consistent with how allied statutes like the Pre-Conception and Pre-Natal Diagnostic Techniques (PCPNDT) Act have been interpreted, is that the medium is irrelevant to the obligation.
Three restrictions matter most for a working marketing team. First, no advertisement may claim or imply a guaranteed pregnancy or a guaranteed live birth outcome — success in ART is probabilistic and clinic-specific, and the Act treats any language suggesting certainty as inherently misleading. Second, no advertisement may offer, suggest, or imply sex selection services in connection with ART, which aligns the Act with the sex-selection prohibitions already in force under the PCPNDT Act. Third, comparative advertising that claims superiority over another named or implied provider is prohibited, which sits close to — but is a distinct statutory basis from — the general comparative-advertising caution under ASCI's Chapter III guidelines that already apply to all healthcare marketing in India.
A fourth, more general obligation under Section 34 prohibits advertising "likely to mislead" a patient about the nature, safety, effectiveness, or probable outcome of an ART procedure — a standard broad enough that ICG treats it as the working default: if a piece of copy could reasonably leave a prospective patient with a false impression of certainty, safety, or comparative advantage, it does not clear the bar, independent of whether it trips one of the three named prohibitions explicitly.
This is a paraphrase of the Act's public text and ICG's working interpretation of it for marketing purposes; it is not a substitute for a clinic's own legal counsel reviewing specific copy against the current statute and any subsequent rules or amendments notified by the Ministry of Health and Family Welfare or the Indian Council of Medical Research (ICMR), which administers the Act.
How it applies to conversational-ad copy specifically
A static Google Search ad or a Meta feed placement has one fixed headline, one fixed description, perhaps three or four rotating variants that a human reviewer can read end to end before the campaign launches. A ChatGPT Ad works differently: the sponsored response is assembled to fit the live conversational turn, drawing from an approved phrase bank and a set of prompt-level constraints, and the exact wording a user sees can vary meaningfully depending on what they just asked the assistant. One user's exploratory "what is IVF success rate at 35" and another's late-stage "book IVF consultation this week" can each surface a differently phrased sponsored response drawn from the same underlying campaign.
That structural difference changes where compliance has to live. Reviewing a single static asset once is not sufficient discipline for a channel that can generate dozens of live phrasings from the same offer. ICG's approach is to push the compliance screen upstream, into the phrase bank and the prompt constraints themselves, rather than trying to review every generated variant after the fact — every candidate phrase that could be assembled into a live ad is screened against the banned-language list before it enters the bank, and the prompt-level constraints explicitly forbid the generation layer from introducing new success-rate claims, comparative language, or outcome guarantees that were not present in the approved source copy.
There is a second, quieter risk specific to the conversational format: a user can push the assistant with follow-up questions inside the same thread — "but what's your actual success rate for my age group" — and a poorly constrained ad system can be nudged into generating a response that drifts from the approved, disclaimer-paired framing into something closer to a bare, unqualified number. ICG's constraint layer caps what the generation system is permitted to surface in response to a follow-up, so a probing question yields a redirect to a qualified statistic and disclaimer, never a bare percentage manufactured on the fly.
The landing page a ChatGPT Ad routes to carries the same obligation as the ad copy itself. A user who clicks through from a conversational ad has already had a partial conversation about outcomes; if the landing page repeats a success-rate figure without the same disclaimer discipline applied in the ad, the compliance gap simply moves one click downstream rather than being closed.
Common violations and how to avoid them
Across client onboarding, the same handful of violation patterns recur in first-draft copy submitted by clinics before ICG's screen runs on it. None of the examples below reference a real clinic — they are composites drawn from patterns ICG has corrected, anonymised.
"Guaranteed pregnancy in your first IVF cycle or your money back." This is the single most common violation ICG encounters. It combines an explicit outcome guarantee, which the Act prohibits outright, with a refund framing that implies the clinic is treating a probabilistic medical outcome as a purchasable, warrantable product — language that would also fail ASCI Chapter III's misleading-claim standard independent of the ART Act specifically.
"70% IVF success rate — book now." A bare percentage with no age band, no denominator, no time period, and critically, no disclaimer that success varies by individual factors, fails the Act's misleading-advertisement standard even though the number itself may be accurate for some subset of the clinic's patients. The number without context reads to a prospective patient as a personal promise.
"Higher success rates than other fertility centres in [city]." Comparative claims against unnamed competitors are still comparative claims under the Act's reading, and ICG treats "other clinics," "leading centres," or any phrasing implying the reader should infer a named alternative is inferior as equivalent risk to naming a competitor outright.
"Family balancing services available" appearing in ART-adjacent ad copy without further qualification. Even where the clinic's actual services are fully compliant, ad copy that could be read as offering sex selection — a euphemism sometimes used in unregulated markets — creates exposure the Act was specifically written to close, and ICG rejects this phrasing outright regardless of the clinic's underlying intent.
The corrective pattern in each case is the same: replace the promise or bare number with a qualified, sourced statistic paired with the mandatory disclaimer, remove any comparative framing entirely, and route every generated ad variant — not just the human-drafted seed copy — through the same screen before it is allowed to serve.
What "clean copy" looks like inside a ChatGPT Ad for this domain
Compliant ChatGPT Ad copy for an IVF or ART client tends to follow a consistent shape: it answers the user's actual question, cites a qualified statistic where relevant, pairs any outcome data with the disclaimer immediately rather than in a footnote, and closes with a low-friction next step rather than a promise.
Query: "what factors affect IVF success rate." Response: "IVF success rates vary by age, ovarian reserve, and underlying diagnosis — clinics report age-banded pregnancy rates per embryo transfer rather than a single figure. [Clinic] is a registered ART clinic under the ART (Regulation) Act 2021; a fertility specialist can review your specific case in a consultation." No number is volunteered unprompted; the disclaimer that success is individual is stated before any figure would even be relevant.
Query: "what is your IVF success rate." Response: "[Clinic]'s reported clinical pregnancy rate per embryo transfer for patients under 35 was [X]% in [year], per internal clinical audit; rates vary meaningfully by age and individual factors and are not guaranteed for any patient. [Clinic] is registered under the ART (Regulation) Act 2021, registration [number]." The figure carries a denominator, an age band, a year, and the disclaimer sits in the same sentence, not a linked footnote.
Query: "book IVF consultation [city] this week." Response: "[Clinic], a registered ART centre in [city], has consultation slots available this week — book a fertility specialist consultation to discuss your specific case and next steps." No outcome claim of any kind appears; the ad converts on availability and registration status, not on a promised result, which is both the compliant framing and, in ICG's testing, a stronger performer on completion rate because it sets an honest expectation before the click.
The pattern generalises: state what varies, cite what is measured with its context attached, name the clinic's Registry status as a trust signal rather than a marketing flourish, and let the call to action be about access — a consultation, a slot, a specialist review — rather than about a result the clinic cannot lawfully promise.
ICG's compliance checklist before every ChatGPT Ad push
Every IVF and ART engagement runs through the same pre-push checklist before a campaign, or any new phrase added to an existing campaign's bank, is allowed to go live. The checklist is logged per push and retained as the client's audit trail, which matters if a regulator or the ad platform itself later asks how a specific piece of copy was approved.
- National Registry status confirmed — the clinic or ART bank's registration under the ART (Regulation) Act 2021 is verified as current before any ART-specific campaign is built, and the registration number is available for inclusion in disclaimer copy.
- Banned-language screen run on every seed phrase — outcome guarantees, bare unqualified percentages, comparative superiority language, and sex-selection-adjacent phrasing are flagged and rejected before entering the approved phrase bank.
- Disclaimer placement verified adjacent to any figure — any success-rate statistic is checked to confirm the variability disclaimer sits in the same unit of copy, not a separate linked page or footnote.
- Generation-layer constraints tested against follow-up probing — ICG runs sample adversarial follow-up questions through the live ad system before launch to confirm it does not drift into unqualified claims under conversational pressure.
- Landing page cross-checked against ad copy — any statistic or claim referenced in the ad is confirmed to appear with matching disclaimer language on the page it routes to.
- Legal-adjacent second read for outcome or comparison language — any copy touching success rates, outcomes, or any comparative framing gets a second reviewer pass beyond the automated screen before approval.
- Surrogacy-adjacent copy routed to a separate track — any ad or landing page touching surrogacy services is reviewed against the Surrogacy (Regulation) Act 2021 separately, since its restrictions are distinct from and stricter than the ART Act's.
This checklist runs alongside, not instead of, ICG's standard healthcare compliance overlay — NMC Section 6, ASCI Chapter III, and DPDP 2023 consent handling — that applies to every healthcare ChatGPT Ads engagement regardless of specialty.
What happens if you're audited
The ART (Regulation) Act 2021 carries meaningful enforcement teeth, and the consequences of a confirmed advertising violation extend well beyond a warning letter. A first contravention of the advertising provisions can attract imprisonment and a fine for the responsible individuals at the ART bank or clinic, with penalties escalating on any repeat violation. Separately and more consequentially for an ongoing business, the National Registry and the state-level appropriate authority administering the Act have the power to suspend or cancel a clinic's registration for a serious or repeated violation — which would end the clinic's legal right to operate as a registered ART provider at all, not merely its right to advertise.
There is also a faster, non-statutory consequence that clients underestimate: the ad platform itself. Google, Meta, and OpenAI's ad network all run their own healthcare-advertising policy reviews independent of Indian statute, and a flagged violation can trigger account-level suspension of ad serving — sometimes within hours of a complaint or automated detection — well before any regulatory process runs its course. In ICG's experience this platform-level suspension is the more immediate business risk for most clients, since it halts the entire campaign's traffic instantly rather than unfolding over the months a statutory enforcement process typically takes.
If an audit or a platform flag does occur, ICG's retained checklist logs become the first line of evidence — showing what screen a specific piece of copy passed, when, and by what constraint set — which materially shortens the time to resolve a platform suspension and gives the clinic's own legal counsel a documented trail to work from if a regulatory inquiry follows. Clients running ART ChatGPT Ads without this kind of audit trail are, in ICG's experience, both more likely to have a violation reach production in the first place and slower to resolve it once flagged.
None of this is a substitute for a clinic's own legal counsel reviewing its specific advertising against the current statute; ICG's checklist is a marketing-compliance discipline built to reduce risk before copy ships, run alongside — not instead of — that legal review.
Running IVF or ART ChatGPT Ads? Get the compliance screen right first.
ICG builds ART Act-clean ChatGPT Ads copy and landing pages for IVF clinics and fertility chains — Registry-checked, disclaimer-paired, audit-logged before anything goes live.
