OSINT Live 2026 Recapped

OSINT Live returned on October 15 for its fifth annual event, bringing investigators and analysts together to explore developments in open-source intelligence (OSINT). This year’s sessions examined how AI is changing investigative research, along with practical techniques for interpreting visual evidence and collaborating with law enforcement. Attendees also got a first look at the public beta of X-Ray, Skopenow’s forthcoming media analysis tool, in a demo that put an AI-manipulated car crash photo to the test. Keep reading for a closer look at what it found.

Accelerating Investigations with AI Agents

Kicking off the day, Tom Vaillant of Buried Signals explored how investigators can coordinate teams of specialized AI agents to expand their research capacity. Dividing the work among agents with different responsibilities can give analysts more time to pursue leads and direct an investigation.

Tom’s approach keeps analysts involved as the research develops. They review the agents’ findings and decide where to go next, retaining responsibility for questioning the results and explaining how they reached their conclusions.

Verifying Visual Evidence

Dr. Manisha Ganguly drew on her reporting for the BBC and The Guardian to explore how open-source investigations can document alleged war crimes when access on the ground is restricted.

Her reconstruction of a Russian detention facility showed how publicly available information, combined with former detainees’ accounts, can reveal conditions inside an otherwise inaccessible location.

Manisha prioritized verifiable evidence over the volume of footage collected. Witness accounts and independent sources helped establish what the imagery showed, while material she could not reliably place within a given timeline stayed out of the reporting.

Investigating Synthetic Identities

Rivka Gewirtz Little, COO of Liminal, examined how synthetic identity fraud is outgrowing the assumptions behind many existing controls. AI is helping fraudsters create convincing digital histories while avoiding certain identifiers that investigators have traditionally used to connect suspicious accounts.

She highlighted the growing role of synthetic identities in opening deposit accounts used to move scam proceeds, as controls built around lending and credit losses may miss identities supporting a much broader fraud operation.

Rivka emphasized the need to look at the infrastructure-related signals behind an identity while also continuing to scrutinize these after onboarding. Last but not least, she cautioned that a sparse history is not proof of fraud and that legitimate customers can have limited records, too.

The Art of Image Analysis

Micah Hoffman and Griffin Glynn of My OSINT Training brought the focus back to observation and the willingness to look at an image differently when an investigation stalls.

Using recreated images from real case scenarios, they showed how seemingly ordinary details can become decisive leads. A hotel room’s predominant color helped identify a location that reverse image searching had missed, while a distinctive brick pattern narrowed the search in another investigation.

Micah and Griffin also showed how documenting observations and tracking completed searches help teams build on each other’s work.

Building Defensible Attribution

Rounding out the day, William Wolfe, formerly a Threat Investigator at Roblox, explored what it takes to turn online findings into intelligence that can support law enforcement.

William recommended corroborating attributions and preserving the evidence behind them. Reports should make clear how investigators reached their conclusions, with the care expected of findings that may be examined in court.

He also discussed the relationships that support collaboration with law enforcement. Investigators need to understand each organization’s information-sharing requirements, with confidentiality and a legitimate investigative purpose guiding the work throughout.

Introducing Skopenow X-Ray

Between sessions, Skopenow previewed X-Ray, which brings AI detection together with metadata and content credential evaluation to help investigators assess media for signs of artificial generation or manipulation.

The demonstration compared a car crash image with an AI-altered version in which the vehicle had been replaced. X-Ray flagged the altered image as likely AI-generated, highlighting inconsistencies and artifacts that supported its finding.

The public X-Ray beta is available to try on our website. X-Ray is not yet commercially available through the Skopenow platform, but stay tuned for updates on the official rollout.

Good OSINT Depends on Good Judgment

Together, this year’s sessions showed why faster collection is only part of OSINT’s development. Whether examining a possible synthetic identity or preparing findings for law enforcement, investigators still need to test the evidence and explain how they reached their conclusions.

Thank you to our speakers and everyone who joined us for OSINT Live 2026. We look forward to seeing you again next year.

Missed the live event? Watch Tom Vaillant’s and Rivka Gewirtz Little’s sessions on demand here.