We disagree with the premise that a confirmed violation under Section 3.18.2 is insufficient to trigger an ADC. The argument that we must first "calibrate" investigative responses based on the "severity" of an indicator fails to account for the unified nature of DNS Abuse and the systemic risk posed by malicious registrants.
An actionable report under Section 3.18.2 is not a mere "indicator"—it is a confirmed breach of the RAA. Once a report is validated, malicious intent is established. To suggest that a domain like bankofamerica1 justifies an ADC while a malware-dropping typosquat like craigslit[.]com does not, creates an arbitrary hierarchy of harm. A registrar cannot claim to have "disrupted" abuse under 3.18.2 if they ignore five other domains in the same account registered with the same malicious signals.
Proportionality must be measured against the severity of the threat to the public. With global cybercrime costs projected by Cybersecurity Ventures to reach $10.5 trillion annually by 2026, the ADC is a necessary and proportionate tool.
Specifically, we must prioritize the fundamental rights of the global public to be secure from financial ruin and identity theft over the procedural "privacy" of a confirmed abuser. Also setting a higher threshold for ADCs than for 3.18.2 mitigation effectively grants professionalized abusers a "free pass" to keep the rest of their malicious infrastructure active.
An ADC is a technical cross-reference, not a content-based inquiry. Because this process is strictly limited to the five technical categories of DNS Abuse, it does not infringe on freedom of expression or lawful speech. It is a technical inquiry into whether other domains are linked to verified malicious activity.
A single, actionable proof of abuse should grant the registrar the latitude—and the obligation—to check associated domains. While registrars should have the discretion to determine the depth of the ADC based on available signals, the initiation of the check must be the default consequence of verified abuse.
We should not use "calibration" as a justification for inaction. The ADC is the proportionate tool required to ensure that the "appropriate mitigation action" mandated by ICANN actually stops the abuse at the source. One actionable proof is enough to turn the key; to do less is to facilitate a $10.5 trillion criminal economy at the expense of global internet users.
-Mary Penn
IPC Representative
_______________________________________________+1, an actionable report pursuant to 3.18.2 should be enough to trigger ADC. After this first step, a registrar should have the latitude to determine the depth and breadth of ADC informed by the signals/indicators available to them and apply mitigation actions where appropriate.
From: Gabriel Andrews via Gnso-dnsabuse-pdp <gnso-dnsabuse-pdp@icann.org>
Reply-To: Gabriel Andrews <gfandrews@fbi.gov>
Date: Wednesday, April 15, 2026 at 5:27 AM
To: "Brian F. Cimbolic" <brian@pir.org>, farzaneh badii <farzaneh.badii@gmail.com>
Cc: "trachtenbergm@gtlaw.com" <trachtenbergm@gtlaw.com>, "volker.greimann@centralnic.com" <volker.greimann@centralnic.com>, "el@lisse.NA" <el@lisse.NA>, "gnso-dnsabuse-pdp@icann.org" <gnso-dnsabuse-pdp@icann.org>, "dns-techs@na-nic.com.na" <dns-techs@na-nic.com.na>
Subject: [EXTERNAL] [Gnso-dnsabuse-pdp] Re: [EXTERNAL EMAIL] - Re: Another numbers request.
Caution: This email originated from outside the organization. Do not click links or open attachments unless you recognize the sender and know the content is safe.
I'm going to speak for a moment on my own behalf to share experience as an investigator, and not on behalf of the GAC.
Farzi, I believe you correctly identify a commonly employed LE principle that the more privacy-invasive an investigative technique is, the more facts/circumstances that may be required to justify its use. Policy is often in place such that a LE investigation can only be opened when there are articulable facts indicating a crime has been committed (for which the agency has authority to investigate). Further, if the investigator wishes to use a particularly privacy-invasive technique, (such as a wiretap to surveil communications in realtime) extensive predication of facts must be presented as to why that level of privacy-invasion is required and couldn't be otherwise satisfied with less privacy invasive techniques.
I believe you incorrectly, however, apply that important principle when you suggest that evidence of a maliciously registered domain - sufficient to trigger 3.18.2 - would not justify taking a look at the other domains registered by that threat actor. This is a minimally invasive investigative step which would be one of the very first steps to take in an investigation once evidence is received of malicious registrations having been made by that customer. It makes use only of information already in possession of the registrar (or reseller), it doesn't piece the veil of protected communications, it's merely a step taken - after you have proof that a domain is maliciously registered - to see what other domains that threat actor is also using maliciously. To not take this step would be, in my view, irresponsible.
Following evidence of malicious registration, an ADC will allow informed mitigative action, which may especially be important to mitigate or prevent victim harm. Example: If a threat actor has registered 100 phishing domains in furtherance of a Business Email Compromise scheme, if a registrar takes piecemeal action only against the one or two domains first reported, the threat actor may yet continue their scheme to defraud the additional 98 victims. Worse, they may accelerate their scheme if already using some of those other 98 domains in communication with victim(s), knowing that the "heat is on". Whereas, if the Rr performed the ADC before taking mitigative action, they may see all 100 domains, and choose to take comprehensive mitigative action against all the phishing domains simultaneously, greatly mitigating potential harm. Recognition of this principle is, I believe, the driving purpose behind this PDP.
All this said - I would greatly benefit from any scenario you might provide in which conducting this simple check might in any way cause harm to an innocent party. I believe Brian asked for such a scenario/example previously, and I have eagerly awaited the same, recognizing that there may be a situation I simply haven't considered.
From: farzaneh badii via Gnso-dnsabuse-pdp <gnso-dnsabuse-pdp@icann.org>
Sent: Wednesday, April 15, 2026 3:44 AM
To: Brian F. Cimbolic <brian@pir.org>
Cc: trachtenbergm@gtlaw.com <trachtenbergm@gtlaw.com>; volker.greimann@centralnic.com <volker.greimann@centralnic.com>; el@lisse.NA <el@lisse.na>; gnso-dnsabuse-pdp@icann.org <gnso-dnsabuse-pdp@icann.org>; dns-techs@na-nic.com.na <dns-techs@na-nic.com.na>
Subject: [EXTERNAL EMAIL] - [Gnso-dnsabuse-pdp] Re: Another numbers request.
I repeat my point, which reflects a global legal practice: the initiation and scope of an investigation must be necessary and proportionate to the available indicators of abuse. We are not confusing investigation and enforcement. Investigative methods themselves, not just penalties, are subject to these requirements. More intrusive or expansive methods must be justified by an initial evidentiary threshold and cannot be used as a default to determine that threshold.
I therefore disagree that ADC is required to assess the nature or extent of an alleged violation. Initial action should be based on indicators derived from the domain itself, such as corroborated abuse reports but also other domain-level signals. We should not be using ADC to determine whether there is sufficient basis to conduct ADC. The appropriate approach is to rely on domain-level indicators first, and only where those establish a sufficient basis, consider more expansive investigative steps, such as examining additional domains associated with an account.
That threshold can be derived from observable indicators of abuse. In phishing cases, for example, the domain string itself can be a strong signal. A domain like “bankofamerica1” may indicate a high likelihood of targeted financial phishing and could justify further scrutiny. By contrast, domains like “youtubee[.]com” or “craigslit[.]com” may suggest typosquatting and potential malware, but those indicators alone do not justify expanding the scope of investigation to associated domains.
The point is that not all indicators justify the same investigative response. The scope of the investigation must be calibrated to the strength and nature of the indicators, and ADC should be reserved for cases where those indicators establish a sufficient basis to expand beyond the domain itself.
Farzaneh
On Wed, Apr 15, 2026 at 1:38 AM Brian F. Cimbolic via Gnso-dnsabuse-pdp <gnso-dnsabuse-pdp@icann.org> wrote:
I agree with Marc - the determination to suspend an individual domain name found during an ADC should rely on the mechanisms already contained in the RAA - if the ADC provides actionable evidence of malicious DNS Abuse, the registrar would be obligated to suspend the relevant domain(s).
If there are 20 other domains in the same registrar account that appear legitimate/benign, I don’t think anyone has ever suggested that those names be suspended too (as there would be no actionable evidence of DNS Abuse under 3.18.2 of the RAA). And if they are, it feels there is consensus that we can nip that line of policy in the bud.
Brian Cimbolic | Chief Legal and Policy Officer
brian@pir.org | www.thenew.org | Power your inspiration. Connect your world.
Confidentiality Note: Proprietary and confidential to Public Interest Registry. If received in error, please inform sender and then delete.
_______________________________________________
Gnso-dnsabuse-pdp mailing list -- gnso-dnsabuse-pdp@icann.org
To unsubscribe send an email to gnso-dnsabuse-pdp-leave@icann.org
Gnso-dnsabuse-pdp mailing list -- gnso-dnsabuse-pdp@icann.org
To unsubscribe send an email to gnso-dnsabuse-pdp-leave@icann.org