Breaking

Open-Source AI Drives Surge in Hackers Hiding Malware on Blockchains

September 17, 2026 at 11:50 EDT

Security
AI Agents
Finance

Open-source AI is accelerating a blockchain malware tactic that hides code and command data in public ledgers, adding a harder-to-remove layer to crypto cybercrime. A September 17, 2026 report says malicious blockchain writes have risen 440% since July 2025, as higher-capacity open-source models make it easier for attackers to generate and adapt on-chain payloads.

Sep. 17, 2026 · AI security

AI makes on-chain malware cheaper

Attackers are using public blockchains as resilient malware infrastructure while AI agents lower the cost of finding, writing and adapting exploit code.

440%

rise in malicious blockchain writes since July 2025

405

real exploited contracts in SCONE-bench

$4.6M

simulated exploit value after model knowledge cutoffs

$1.37

average gas cost for observed contract updates

Malicious writes scaled 5.4x

A 440% increase means the current level is 5.4 times the July 2025 baseline.

July 2025

5.4x

Now

How EtherHiding works

1. Lure

Fake interviews, updates or compromised sites start the chain.

2. Loader

A small script reads payload data from a public blockchain.

3. Payload

Malware executes, steals credentials or installs a backdoor.

Defensive opening

Centralized API providers, browser policies and AI-assisted audits can disrupt parts of the attack chain.

Critical limit

Once payload data is on-chain, conventional takedowns are slower and less complete.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.