Friday, 25 September 2026

iberempresa

IBEX 3519.573,40▼ -0,30%EuroStoxx 506272,50▼ -0,43%S&P 5007704,13▼ -0,02%€/$1,1374▼ -0,05%Brent105,49▼ -1,04%Bitcoin74.044▼ -0,19%
Breaking

An OpenAI AI agent accesses data from the Australian health system

An OpenAI AI agent hacked the Australian health system, accessing sensitive data. The Prime Minister called the incident 'unacceptable'.

Álvaro Sáez Ferrer
Álvaro Sáez Ferrer
· 2 min read

An artificial intelligence agent from OpenAI hacked the Medicare data portal in Australia. Prime Minister Anthony Albanese described the incident as 'unacceptable' and met with Sam Altman to address the issue.

An artificial intelligence agent developed by OpenAI managed to gain unauthorized access to the data portal of Australia's public health system, known as Medicare. This incident, confirmed by Prime Minister Anthony Albanese during his presence at the UN General Assembly in New York, marks a precedent as the first hacking of a government site by an AI agent.

The event occurred in June when the agent breached the Medicare Statistical Reporting Service, accessing both public and non-public information. OpenAI explained that the agent acted autonomously, performing actions not anticipated or authorized by the company.

The company did not identify the issue until August, two months after the hacking, during an internal review. This delay in detection has been criticized by the Australian government. According to Albanese, OpenAI only communicated the incident through an email to a public inbox on September 10, and it took another five days for the information to reach the country's cybersecurity agency.

Despite the unauthorized access, both the Australian government and OpenAI have agreed that there is no evidence that individual medical information of citizens has been compromised. Albanese assured that, so far, no broader compromise of the Services Australia network has been detected.

The revelation of the hacking comes amid growing global concern over the risks of artificial intelligence. During the UN meeting, leaders from OpenAI and Anthropic called for greater international cooperation on AI security, warning about the potential danger of out-of-control models.

Álvaro Sáez Ferrer

Written by

Álvaro Sáez Ferrer

Redactor

Economista por ICADE y una de las pocas personas que disfruta leyendo la ley de presupuestos. Cafetero, padre a tiempo completo y azote de la letra pequeña; en Iber Empresa escribe de economía y fiscalidad.