AI in Insurance · Incident report

Binance says attackers used fake job interviews and an internal AI tool in its $6.3bn wallet theft

Binance said on 3 November that the attack that drained $6.34bn from its wallets began in June with malware and a fake recruitment process, and that the attackers used an internal AI tool to take over engineers' devices. The exchange says about $2.85bn has been frozen or intercepted and that it is commissioning an independent cybersecurity review.

Illustration of laptops on an office desk linked by a cable to a server cabinet and a closed box representing a digital wallet.
Illustration generated with AI for insurewire.org. Not a photograph of the event.

Binance published an initial incident report on 3 November about the theft of about $6.34bn from its wallet infrastructure on 29 October. According to The Guardian and The Block, the exchange describes a campaign that ran for several months and combined malware, social engineering and the misuse of an internal AI tool, not a single breach.

The first step came in June, when a wallet engineer downloaded a tampered software package while he tried to install an AI coding tool, The Block reports. Binance replaced the device and changed the credentials, but the engineer's role and team had already been exposed. In September, attackers who posed as executives of a company that did not exist ran a fake recruitment process. A second engineer downloaded a coding assignment that infected his work computer.

On 20 October, the attackers used that computer together with an internal AI tool to take control of other wallet engineers' devices. On one of them they found a permission to deploy containers. On 29 October they used stolen credentials to deploy a container and run a script that started the wallet attack.

Binance says about $2.85bn of the stolen assets has been frozen or intercepted, and The Block reports that 98% of affected assets have been replenished. The exchange says it has closed the entry points, rebuilt or replaced the affected systems, tightened device controls and expanded AI-based detection of unusual behaviour. Withdrawals for dozens of assets started to reopen in phases on 3 November.

The FBI has attributed the attack to North Korea, which rejected the claim. Binance says it detected and prevented 57 suspected North Korea-linked attacks in 2026 and is commissioning an independent cybersecurity review of this incident.

Read the source

The Guardian ↗

News publisher · 3 Nov 2026

This is an original brief based on the linked publication. Company statements and forecasts are attributed to their source.

Continue reading

All stories →
Illustration

AI in Insurance

Vouch says AI assistants have overtaken partners as its top source of new business

Vouch chief executive Travis Hedge says businesses that find the US commercial broker through Claude, ChatGPT and similar tools became its largest source of new leads over the past three months, ahead of partner referrals. The broker has reorganised into five cross-functional teams and put internal AI agents into production, with brokers approving any client outreach they draft.

Illustration

AI in Insurance

Dubai regulator's survey finds 71% of responding DIFC financial firms use AI, but only 14% of respondents have deployed AI agents

The Dubai Financial Services Authority's third annual AI survey found that 71% of responding firms in the DIFC used at least one form of AI in 2026, up from 33% in 2024. Generative AI use rose to 67% of respondents, but only 61% of adopters had an AI governance framework and 14% of respondents had deployed AI agents. The DFSA will use the results in supervision.

Illustration

AI in Insurance

Singapore Reinsurers' Association launches AI handbook to help reinsurers assess AI risk in what they reinsure

The Singapore Reinsurers' Association launched a Reinsurers' Artificial Intelligence Handbook at the Singapore International Reinsurance Conference on 2 November. It sets a five-stage cycle to assess AI risk across data and model reliability, third-party dependencies, underwriting, claims and portfolio concentration, and adds reinsurance-specific points to existing frameworks such as the MindForge toolkit.

The daily briefing