THN Cybersecurity Recap: Latest News, Tools and Tips (Oct 21 - Oct 27)

Delta Air Lines Sues CrowdStrike for July Outage: Delta Air Lines filed a lawsuit against CrowdStrike in the U.S. state of Georgia, accusing the cybersecurity vendor of breach of contract and negligence after a major outage in July caused 7,000 flight cancellations, disrupted travel plans of 1.3 million customers, and cost the carrier over $500 million. "CrowdStrike caused a global catastrophe because it cut corners, took shortcuts, and circumvented the very testing and certification processes it advertised, for its own benefit and profit," it said. "If CrowdStrike had tested the Faulty Update on even one computer before deployment, the computer would have crashed." CrowdStrike said "Delta's claims are based on disproven misinformation, demonstrate a lack of understanding of how modern cybersecurity works, and reflect a desperate attempt to shift blame for its slow recovery away from its failure to modernize its antiquated IT infrastructure."

Meta Announces Secure Way to Store WhatsApp Contacts: Meta has announced a new encrypted storage system for WhatsApp contacts called  Identity Proof Linked Storage (IPLS), allowing users to create and save contacts along with their usernames directly within the messaging platform by leveraging key transparency and hardware security module (HSM). Until now, WhatsApp relied on a phone's contact book for syncing purposes. NCC Group, which carried out a security assessment of the new framework and uncovered 13 issues, said IPLS "aims to store a WhatsApp user's in-app contacts on WhatsApp servers in a privacy-friendly way" and that "WhatsApp servers do not have visibility into the content of a user's contact metadata." All the identified shortcomings have been fully fixed as of September 2024.

CISA, FBI Investigating Salt Typhoon Attacks: The U.S. Cybersecurity and Infrastructure Security Agency (CISA) said the U.S. government is investigating "the unauthorized access to commercial telecommunications infrastructure" by threat actors linked to China. The development comes amid reports that the Salt Typhoon hacking group broke into the networks of AT&T, Verizon, and Lumen. The affected companies have been notified after the "malicious activity" was identified, CISA said. The breadth of the campaign and the nature of information compromised, if any, is unclear. Recent reports from The New York Times, The Wall Street Journal, Reuters, and CBS News have claimed that Salt Typhoon used their access to telecommunications giants to tap into phones or networks used by Democratic and Republican presidential campaigns.

Fraudulent IT Worker Scheme Becomes a Bigger Problem: While North Korea has been in the news recently for its attempts to gain employment at Western companies, and even demanding ransom in some cases, a new report from identity security company HYPR shows that the employee fraud scheme isn't just limited to the country. The company said it recently offered a contract to a software engineer claiming to be from Eastern Europe. But subsequent onboarding and video verification process raised a number of red flags about their true identity and location, prompting the unnamed individual to pursue another opportunity. There is currently no evidence tying the fraudulent hire to North Korea, and it's not clear what they were after. "Implement a multi-factor verification process to tie real world identity to the digital identity during the provisioning process," HYPR said. "Video-based verification is a critical identity control, and not just at onboarding."

Novel Attacks on AI Tools: Researchers have uncovered a way to manipulate digital watermarks generated by AWS Bedrock Titan Image Generator, making it possible for threat actors to not only apply watermarks to any image, but also remove watermarks from images generated by the tool. The issue has been patched by AWS as of September 13, 2024. The development also follows the discovery of prompt injection flaws in Google Gemini for Workspace, allowing the AI assistant to produce misleading or unintended responses, and even distribute malicious documents and emails to target accounts when users ask for content related to their email messages or document summaries. New research has also found a form of LLM hijacking attack wherein threat actors are capitalizing on exposed AWS credentials to interact with large language models (LLMs) available on Bedrock, in one instance using them to fuel a Sexual Roleplaying chat application that jailbreaks the AI model to "accept and respond with content that would normally be blocked" by it. Earlier this year, Sysdig detailed a similar campaign called LLMjacking that employs stolen cloud credentials to target LLM services with the goal of selling the access to other threat actors. But in an interesting twist, attackers are now also attempting to use the stolen cloud credentials to enable the models, instead of just abusing those that were already available.