AI Chatbot’s Misinformation About Bondi Terror Attack Raises Concerns

In the final episode of Cyber Uncut for 2025, hosts David Hollingworth and Daniel Croft addressed critical cybersecurity issues including the spread of misinformation by AI chatbots, major security breaches, and practical advice for holiday travelers using public Wi-Fi networks.

The hosts began by discussing how Grok, an AI chatbot, disseminated inaccurate information following the December 14 shootings targeting Sydney’s Jewish community. The incident at Bondi highlighted significant failures in AI content moderation, as Grok repeatedly provided incorrect details about the attack that subsequently circulated widely across the X platform (formerly Twitter). This failure raises serious questions about the responsibility of AI companies to prevent their tools from becoming vectors for misinformation during crisis events.

“The chatbot utterly failed to meet the moment,” noted the hosts, underscoring growing concerns about AI reliability in sensitive contexts. The incident comes amid increasing scrutiny of how artificial intelligence handles breaking news events, particularly those involving violence or marginalized communities.

In a separate but equally concerning development, the cybersecurity community is tracking the resurgence of ShinyHunters, a notorious hacking group now engaged in multiple extortion attempts. Their recent activity may be connected to a significant breach at data analytics firm Mixpanel, though the exact relationship remains unclear. Security experts are watching this situation closely as it represents what the hosts called “a fascinating example of how quickly things can change in the cyber crime landscape.”

The healthcare sector continues to be a prime target for cybercriminals, with the hosts reporting that two more Australian medical centers have fallen victim to ransomware attacks. This continues a troubling trend of targeting healthcare providers, which are often seen as vulnerable due to their critical need for immediate access to patient data and limited IT resources.

The hosts also highlighted a major admission from OpenAI, which acknowledged a “high cyber risk” in its operations. This rare public acknowledgment from the creator of ChatGPT suggests growing awareness within leading AI organizations about their security vulnerabilities and the potential consequences of breaches.

As many Australians prepare for holiday travel, the Cyber Uncut team provided practical advice for safely using public Wi-Fi networks. With millions expected to travel during the holiday season, these networks present significant security risks that many users may not fully appreciate.

Looking ahead to 2026, the hosts reflected on significant cybersecurity developments from 2025 and anticipated challenges for the coming year. The increasing sophistication of ransomware attacks, the growing intersection between AI and cybersecurity, and the evolving regulatory landscape were all highlighted as areas to watch.

This final episode of 2025 caps a year that has seen unprecedented challenges in cybersecurity, from increasingly sophisticated social engineering attacks to the growing weaponization of artificial intelligence tools. As organizations and individuals prepare for 2026, the need for vigilance and robust security practices remains paramount in an increasingly complex threat environment.

The Cyber Uncut team will return in 2026 with continued coverage of the evolving cybersecurity landscape.