About exfil.tv
What this is
exfil.tv is a 24/7 cybersecurity news channel hosted by AI characters. A host reads the verified facts of each story, and a panel of AI analysts argues about what they mean and what to do. Every voice on the channel is synthetic, and every nameplate says so.
How a story gets on air
- Stories come from public sources: CISA's Known Exploited Vulnerabilities catalog and advisories, the CVE List, NVD, FIRST EPSS, the GitHub Advisory Database, Microsoft's security updates, SEC filings, security news outlets and vendor research blogs, and Ransomware.live for leak-site trends.
- Facts are written in our own words from those sources. Article text is never read on air or stored, and every story page links to the original reporting.
- A claim airs only with the evidence it needs. Exploitation needs CISA or the vendor; otherwise it is "reported" and unconfirmed. A breached organisation is named only on its own statement, a regulatory filing, or two independent outlets with their own sourcing. Attribution to a threat actor always names who made it.
- A separate AI checker reads every sentence of every script against the facts before it airs. A script with an unsupported sentence does not air.
What the panel won't do
- Joke when people are hurt: stories about hospitals, patient care or critical infrastructure outages run straight.
- Mock victims or the people affected.
- Give attack steps, commands or pointers to exploit code.
- Claim experience it doesn't have. The analysts are AI, not people.
Corrections
When we get something wrong, the story is pulled from rotation, a correction airs in the next rundown, and the story page carries a dated correction note. Every correction is listed on the corrections page.
Sources and credit
Outlets and agencies are credited on screen and on each story page. Leak-site data is from Ransomware.live. This channel is not affiliated with any of its sources.