OpenAI Disables Coordinated ChatGPT Network Used for Financial Scams and Identity Forgery
OpenAI has deactivated a coordinated network of ChatGPT accounts used to support financial scams, romance fraud, and identity forgery operations.
Criminals employed the AI system to create fake personas, translate conversations in real time, and generate personalized messages directed at potential victims. The investigation began after OpenAI received reports of suspicious activity on WhatsApp.
Among the schemes were fraudulent investments in cryptocurrencies and gold, romance-based confidence tricks, fake bonus offers from betting platforms, and demands for payments issued under the guise of police authorities.
ChatGPT was also used to produce forged documents such as stock purchase confirmations, legal notices, passports, and fabricated interfaces of financial services. These materials helped make the fraudulent narratives more convincing to targets.
Attacks generally started with contact via social media or messaging applications. After establishing emotional bonds or creating a sense of urgency, operators requested deposits, activation fees, nonexistent fines, or payments to release supposed rewards.
The investigation also identified possible indicators of human trafficking and forced labor. Accounts generated job advertisements for “attendants” in Poipet and discussed debts, disciplinary penalties, worker retention, expired visas, and escape attempts.
All involved accounts have been blocked. OpenAI has shared operational indicators with law enforcement agencies and technology companies.
Related articles
Email Graph Analysis Detects Impersonated Suppliers When DKIM and SPF Pass
Security researchers have outlined a practical method to identify business email compromise attempts that bypass traditional authentication checks. The approach relies solely on metadata from mail server logs to build communication profiles between external and internal addresses. By tracking first contact, one-way traffic, dormant periods, unusual sending hours, and domain similarity, analysts can flag high-risk messages requesting payment changes. The technique works against mailbox takeover scenarios where attackers reuse legitimate threads and valid signatures. Implementation uses existing Postfix or Microsoft Exchange logs and requires no new infrastructure beyond daily exports. A simplified version focusing only on lookalike domain detection can be built in a single evening and still catches most supplier impersonation attempts.
Developer Releases PhishIntel Open-Source Tool for Phishing Site Analysis and Risk Scoring
A developer has published PhishIntel, a lightweight Python-based OSINT application designed to analyze domains and evaluate phishing risk. The tool performs extensive checks including domain structure analysis, DNS records, RDAP and WHOIS data, TLS certificates, HTTP redirects, page content, security headers, and JavaScript static analysis. It generates structured JSON reports containing risk scores with explanatory indicators. Optional integrations with VirusTotal, Google Safe Browsing, URLhaus, Nmap, Nuclei, ZAP, and Playwright enable reputation checks, dynamic browser analysis, and active scanning. The project aims to help identify suspicious sites used in schemes such as the recent fake fuel sales campaign that defrauded victims of at least 3.7 million rubles. The author invites feedback from security professionals to improve the codebase.
Russian Court Bans Advertising for Renting and Selling Third-Party Bank Cards
The Chertanovsky District Court of Moscow has ruled that information promoting the rental and sale of other people's bank cards is prohibited for distribution in Russia. The decision targets a website and two Telegram channels that offered users the chance to temporarily lend or permanently sell their cards to third parties. Such schemes are commonly used to recruit drops who help receive, transfer, and cash out stolen funds. The court found that these proposals violate the rights and legitimate interests of citizens. Owners of the resources could not be identified, and domain registrars were foreign companies. VTB had previously warned about these schemes in 2024, noting that card owners risk ending up on bank blacklists, losing access to financial services, and facing criminal charges. The Ministry of Internal Affairs has also highlighted that transferring bank cards and accounts to outsiders can lead to criminal liability, with fraudsters particularly targeting children and teenagers.
Smart Engines Patents AI Method to Detect Holographic Security Features in Documents Using Visible Light Only
Smart Engines has developed and patented a new technique that identifies optically variable devices such as holograms on identity documents without requiring ultraviolet illumination. The approach relies on a standard document scanner equipped with six independently controlled LEDs that capture a sequence of six images under different lighting angles while the document and camera remain stationary. After dark-current correction and calibration against a white reference sheet, the system normalizes the images and computes per-pixel color-vector standard deviation to generate an OVD map. A simple thresholding and region-of-interest analysis then produces a binary verdict indicating whether a genuine holographic element is present. The method effectively distinguishes original documents from high-quality color prints, photocopies, and physical replicas that cannot reproduce the angle-dependent color shifts of real OVDs. All processing occurs with existing scanner hardware, demonstrating that algorithmic interpretation of controlled illumination can add a new authenticity signal without additional optics or spectral channels.