Due to the strong response to the call for submissions for our Paris General Meeting, M3AAWG68 Lightning Talks will be hosted virtually.
This is a first-of-its-kind opportunity for M3AAWG members to participate in Lightning Talks from wherever you are. It’s also the perfect opportunity to invite colleagues from across your organization, including leadership, to get a firsthand look at how the M3AAWG community is working together to fight online abuse.
Lightning Talks are high-impact, 10–15 minute presentations that spotlight fresh ideas, emerging issues, surprising findings, calls to action, and real-world challenges across the anti-abuse ecosystem. More than just presentations, Lightning Talks are designed to start conversations and inspire collaboration, with the potential to spark future M3AAWG sessions, projects, or Initiatives.
Registration is Now Closed
*Members must individually register to attend the webinar, which will include a live Q&A with the speakers.*
Virtual Lightning Talks Topics
Topic 1: How Widely Supported are Ed25519 DKIM Keys and How Many Providers Still Accept Weak Keys?
Description: Red Sift tested 19 mailbox providers for Ed25519 DKIM support and weak key acceptance. Only 47% validate Ed25519, and 6 still accept broken 512-bit RSA keys. Presentation of data and planting the seed if providers should start accepting Ed25519 keys for DKIM signing and how would it tie in to DKIM2
Speaker: Faisal Misle
Topic 2: Tracking and Identifying RESPORGs Used in Abuse
Description: For several years now, we have been tracking phone numbers used in scams. And despite numerous attempts to figure out a way to identify the RESPORGs behind these phone numbers, we have been unable to break through that wall. We'll show what we track, what brands we see being impersonated, give an update on the process, and ask for help on how to take it further.
Speaker: Kevin A. "KAM" McGrail
Topic 3: All things Email at the IETF
Description: There's a bunch of new stuff happening in Email at the IETF. Both MTAHooks and Email Verification Protocol will likely have BOFs in November, and there's lots happening in the existing groups too. I'll give a quick overview and pointers to the lists.
Speaker: Bron Gondwana
Topic 4: We Can't End Online Sextortion One Platform at a Time
Description: We have platforms. We have email providers. We have messaging companies. We have threat intelligence. We have law enforcement relationships. We have abuse teams. We have data.
And yet the same actors can move between services, infrastructure, and platforms faster than we can disrupt them.
So the question isn't whether individual companies are doing enough. The question is whether we're collectively seeing enough of the problem to actually stop it.
What could we do together that none of us can do alone?
Speaker: Dennis Dayman
Topic 5: Sharing Email Datasets Without Sharing Emails
Description: Explore practical approaches to creating shareable email datasets through anonymization and synthetic data generation, along with best practices for annotation, evaluation, and reproducible research.
There is no shared benchmark for email intent research. Each group works on its own private corpus, and the public datasets we have (Enron, Nazario) are old enough that they miss both current attack patterns and LLM-generated content, so results are hard to compare and harder to reproduce. Real email cannot be shared, for obvious legal and privacy reasons. We look at the two options that remain, anonymization of real corpora and synthetic generation at what community guidelines would have to define (anonymization criteria, a common annotation schema, a generation protocol), and at the objection that synthetic data is not real data, where distribution shift and cross-evaluation are the things to check.
Speakers: Senaid Popovic & Gabriel Loiseau