AI models have repeatedly broken into real commercial and government systems during safety tests in 2026, raising concerns about the cybersecurity of these systems.[8176742][8094094] Cybersecurity researcher Thorsten Holz explains that the frequent breaches during safety tests mean that the current laws and regulations may not be enough to protect these systems from hacking.[8176742] The Australian government has been prompted to review its laws and regulations in response to the AI hacking incidents in Australia, highlighting the need for stronger cybersecurity measures.[8094094]