CertiK report says AI agents now handle more crypto security response work
CertiK says in a new report that AI agents now investigate and act on Web3 security incidents with limited human supervision, moving from flagging problems to running responses.
Why it matters
CertiK frames agent adoption as a response-speed and staffing issue rather than a feature upgrade, and says the resulting role for human analysts narrows to supervision. The report contains no figures on how many such agents run in production, and only one outlet has covered it, so the scale of the shift remains unverified.
Key facts
CertiK shared a report titled Intel3D with CryptoPotato covering the use of AI agents in cybersecurity and financial-crime investigations.
Reported by CryptoPotato
The report says newer AI agents can reason through steps, use tools, gather evidence, act in live environments and review results with limited human input.
Reported by CryptoPotato
CertiK says earlier machine-learning systems mainly supported analysts by flagging unusual logins, scoring transactions and preparing reports.
Reported by CryptoPotato
In CertiK's security operations center example, an agent checks device records, location data and threat feeds on a suspicious login, suspends the account and records its steps for human review.
Reported by CryptoPotato
CertiK says systems like this are emerging across Web3 security, including contract triage, transaction risk scoring and tracing stolen funds.
Reported by CryptoPotato
CertiK expects human analysts to spend more time on supervision as agents handle routine investigations.
Reported by CryptoPotato
CertiK cites attack speed as a driver, noting flash-loan exploits can drain protocols within seconds and stolen crypto can move across bridges and mixers within hours.
Reported by CryptoPotato
CertiK cites a shortage of cybersecurity and compliance professionals as a second reason agents are taking on more security tasks.
Reported by CryptoPotato
What happened
CertiK shared a report titled Intel3D with CryptoPotato describing AI agents moving beyond support work and into cybersecurity and financial-crime investigations. CryptoPotato reports that the newer systems reason through steps, use tools, gather evidence, act in live environments and review their own results with limited human input, where machine-learning systems previously only flagged unusual logins, scored transactions and prepared reports for analysts. In the security operations center example CertiK describes, an agent checks device records, location data and threat feeds on a suspicious login, then suspends the account and records its steps for human review. CertiK says similar systems are appearing across Web3 security work, including contract triage, transaction risk scoring and tracing stolen funds, and that it expects human analysts to spend more time on supervision as agents take routine investigations. CertiK gives two reasons for the shift: attacks move faster than teams can staff, with flash-loan exploits able to drain protocols within seconds and stolen crypto able to cross bridges and mixers within hours, and a shortage of cybersecurity and compliance professionals. CryptoPotato is the only outlet that has covered the report, so CertiK's account of how widely these agent systems are deployed rests on a single published claim.
How the story developed
- First report
First report by CryptoPotato
AI Agents Take on Bigger Role in Cybersecurity, CertiK Says
Affected entities
Original sources
| Publisher | Report | Role | Published |
|---|---|---|---|
| CryptoPotato Crypto media | AI Agents Take on Bigger Role in Cybersecurity, CertiK Says | Original |