Azerbaijan’s Deputy Foreign Minister Elnur Mammadov says AI is speeding up cyberattacks and making deepfakes more convincing. At a cyber diplomacy conference in Baku, he called for faster cross-border warnings and set out three practical priorities for governments.
Watch Desk analysis
What happened
Mammadov said AI-generated voice messages were used in a June 2025 impersonation campaign targeting three foreign ministers. He also cited CrowdStrike’s 2026 Global Threat Report, which says attackers in observed cybercrime intrusions in 2025 took an average of 29 minutes to move beyond the system they first compromised; the fastest recorded instance took 27 seconds. Those figures are attributed to the report, not independently assessed here.
His proposed response is to establish secure communication channels before a crisis, agree in advance what threat information can be shared and with whom, and build trust through exercises, expert exchanges and engagement with technology companies. He argued that early warnings should be shareable before responsibility for an attack is conclusively established, while clearly marking what remains uncertain. Trend.Az reported his remarks from the conference.
Why it matters
That is a useful distinction: warning partners that a threat exists need not mean claiming to know who is behind it. If verification procedures take longer than an attack takes to spread, waiting for a finished investigation can leave defenders with a very tidy account of what they might have stopped.
Mammadov also pointed to the EU AI Act’s transparency and labelling provisions for AI-generated content, including deepfakes, and called for cooperation and capacity-building across countries. His comments are a policy argument, not evidence that these measures have been adopted as a new international system.
Our read
The strongest part is the emphasis on arrangements made before the alarm goes off: tested contacts, agreed sharing rules and regular cooperation. AI-generated impersonation is not just a content-moderation headache when false statements can muddy diplomatic communications. The test is whether governments can share a useful warning quickly without turning uncertainty into accusation.
What to watch
- Whether governments or incident-response teams agree on secure, tested channels for early warnings.
- Whether information-sharing rules allow rapid alerts while making uncertainty explicit.
- Whether the conference produces concrete follow-up, including exercises or cross-border agreements.
Discussion spark: Should governments share cyber threat warnings before they know who is responsible, even if that means circulating claims that may later prove wrong?
Sources and evidence
- AI accelerates cyberattacks, making faster information sharing essential – Deputy FM – Trend.Az (19 November 2026, 07:44 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.