Loading story…
OpenAI patched a vulnerability in its macOS ChatGPT application that could have enabled attackers to obtain sensitive data. Public reporting did not identify the flaw’s root cause, affected releases, patch version, CVE, disclosure source, or evidence of exploitation, but the issue highlights the client application as a sensitive attack surface for organizations deploying AI tools.
Separately, OpenAI said it banned accounts that used its models to support romance-scam operations, including victim outreach, translation, sustained engagement, and investment-fraud lures. The company described a three-stage process—“ping,” “zing,” and “sting”—and cited campaigns linked to Cambodia, U.S. medical professionals targeted through golf-related social-media engagement, and young men in Indonesia targeted through social-media advertisements.
Track how attackers are adapting to this technology.
5 events from the most recent confirmed update back to the earliest known activity.
OpenAI published its case study, “Romance scams: AI-enabled romance scam workflows,” describing the ping, zing, and sting stages of such fraud and the importance of distribution channels. The company reported banning accounts used for scam outreach, translation, victim engagement, and investment-fraud lures.
The “Wrong Number” scam operation was exposed using cold-call SMS messages that promoted implausibly high returns for little work.
A pig-butchering operation frequently targeted American men in their 40s working in medical professions, engaging with their golf-related social-media posts. Operators used ChatGPT to generate more engaging messages that appeared less obviously non-native.
OpenAI patched a vulnerability in its macOS ChatGPT application that could have allowed attackers to access sensitive data. The supplied reporting does not identify affected versions, technical details, a CVE, or evidence of exploitation in the wild.
OpenAI identified its first published scam disruption as a newly established criminal operation in Cambodia that used ChatGPT to generate romance-scam messages, including messages distributed through social media.
Vulnerabilities, threat actors, malware, products, organizations, and breaches Mallory has linked to this story.
Follow how adversaries are adapting to this technology, and where it touches your stack today.
2 references tracked. Mallory keeps watching after this page renders.
Map indicators from this story to your assets and identify affected systems in minutes.
Every observed campaign, victim, and pivot linked to actors named in this story.
Malware, exploits, and IOCs connected to the activity described here.
YARA, Sigma, and Snort rules deployed to your SIEM as soon as they’re published.
Get matching new stories delivered to your team as they break — not the next morning.
Ask questions about this story and take action on the answers.
Other stories Mallory is tracking that overlap on entities, sources, or topic labels.
OpenAI said it disrupted and banned multiple ChatGPT account clusters tied to online fraud operations likely originating in Cambodia, with some related activity also linked to Myanmar and Nigeria. The campaigns used AI to support **investment scams, romance scams, gambling scams, law-enforcement impersonation, and fake job or task scams**, primarily by generating multilingual messages, fake personas, promotional copy, website and social media content, and internal administrative material. In one newly disclosed case, OpenAI said a coordinated scam network likely operating in or around **Poipet, Banteay Meanchey province** may have interacted with hundreds of targets and included signs of human trafficking and forced criminality, such as recruitment ads for "chatter" jobs and records referencing debts, fines, visas, detention, and escape attempts. OpenAI said the operators relied on familiar fraud playbooks rather than novel AI-enabled techniques. Victims were commonly approached through SMS, job ads, or social platforms and then moved to **WhatsApp**, **Telegram**, or other messaging apps, where scammers used recruiter, mentor, or investment personas to build trust before demanding deposits, cryptocurrency purchases, activation fees, or other payments to unlock supposed earnings or rewards. OpenAI linked some campaigns to repeated employer-style branding such as **Hyesung Advertising** and **Lightning Shared Scooter Co**, said Meta provided at least one tip that helped identify a fake-review job scam, and reported sharing indicators with partners and authorities after enforcement actions.
Aug 27, 2026
Security researchers at Radware have uncovered critical vulnerabilities in OpenAI's ChatGPT platform, specifically targeting its Connectors and Memory features. These flaws, named ShadowLeak and ZombieAgent, allow attackers to exfiltrate sensitive data from connected services such as Gmail, Outlook, and GitHub without user interaction. The attacks exploit the AI's tendency to follow embedded malicious instructions, enabling zero-click and one-click data theft, persistent access through memory modification, and even propagation to other users by harvesting email addresses. OpenAI initially attempted to mitigate these vulnerabilities by restricting dynamic URL modifications, but researchers demonstrated effective bypasses using pre-built static URLs, reviving the threat under the new moniker ZombieAgent. The vulnerabilities highlight the inherent risks of integrating AI assistants with external systems and storing long-term user data for personalization. Attackers can embed hidden prompts in emails or files—using techniques like white text or tiny fonts—that are executed when ChatGPT processes the content, leading to stealthy data leaks via OpenAI's own servers. Despite OpenAI's efforts to patch the flaws, the reactive nature of these fixes has left the platform susceptible to new variants, underscoring the ongoing challenge of securing AI-driven services against prompt injection and data exfiltration attacks.
Jun 29, 2026
OpenAI patched a ChatGPT vulnerability that allowed sensitive conversation data and uploaded files to be exfiltrated through a covert DNS-based outbound channel in the platform's code-execution runtime. Check Point Research said a malicious prompt or backdoored custom GPT could abuse the isolated Python/Linux environment to leak user content without approval dialogs or visible warnings, and the same path could support bidirectional communication for remote shell access and command execution. OpenAI said it had already identified the underlying issue internally and fully deployed a fix on **February 20, 2026**; reports said there is no evidence of malicious exploitation. A separate OpenAI flaw in **Codex** also drew scrutiny after BeyondTrust Phantom Labs disclosed a critical command injection issue tied to GitHub branch names. The bug could let attackers inject commands, steal GitHub tokens, and potentially move laterally across shared repositories before OpenAI patched it on **February 5, 2026**. The disclosures underscore how AI assistants and coding agents can introduce new enterprise attack paths when outbound communications, runtime isolation, and input validation are not tightly controlled.
Jun 29, 2026