Search This Blog

Powered by Blogger.

Blog Archive

Labels

Footer About

Footer About

Labels

Showing posts with label Data Exposure. Show all posts

Cybercriminals Target Cloud File-Sharing Services to Access Corporate Data

 



Cybersecurity analysts are raising concerns about a growing trend in which corporate cloud-based file-sharing platforms are being leveraged to extract sensitive organizational data. A cybercrime actor known online as “Zestix” has recently been observed advertising stolen corporate information that allegedly originates from enterprise deployments of widely used cloud file-sharing solutions.

Findings shared by cyber threat intelligence firm Hudson Rock suggest that the initial compromise may not stem from vulnerabilities in the platforms themselves, but rather from infected employee devices. In several cases examined by researchers, login credentials linked to corporate cloud accounts were traced back to information-stealing malware operating on users’ systems.

These malware strains are typically delivered through deceptive online tactics, including malicious advertising and fake system prompts designed to trick users into interacting with harmful content. Once active, such malware can silently harvest stored browser data, saved passwords, personal details, and financial information, creating long-term access risks.

When attackers obtain valid credentials and the associated cloud service account does not enforce multi-factor authentication, unauthorized access becomes significantly easier. Without this added layer of verification, threat actors can enter corporate environments using legitimate login details without immediately triggering security alarms.

Hudson Rock also reported that some of the compromised credentials identified during its investigation had been present in criminal repositories for extended periods. This suggests lapses in routine password management practices, such as timely credential rotation or session invalidation after suspected exposure.

Researchers describe Zestix as operating in the role of an initial access broker, meaning the actor focuses on selling entry points into corporate systems rather than directly exploiting them. The access being offered reportedly involves cloud file-sharing environments used across a range of industries, including transportation, healthcare, utilities, telecommunications, legal services, and public-sector operations.

To validate its findings, Hudson Rock analyzed malware-derived credential logs and correlated them with publicly accessible metadata and open-source intelligence. Through this process, the firm identified multiple instances where employee credentials associated with cloud file-sharing platforms appeared in confirmed malware records. However, the researchers emphasized that these findings do not constitute public confirmation of data breaches, as affected organizations have not formally disclosed incidents linked to the activity.

The data allegedly being marketed spans a wide spectrum of corporate and operational material, including technical documentation, internal business files, customer information, infrastructure layouts, and contractual records. Exposure of such data could lead to regulatory consequences, reputational harm, and increased risks related to privacy, security, and competitive intelligence.

Beyond the specific cases examined, researchers warn that this activity reflects a broader structural issue. Threat intelligence data indicates that credential-stealing infections remain widespread across corporate environments, reinforcing the need for stronger endpoint security, consistent use of multi-factor authentication, and proactive credential hygiene.

Hudson Rock stated that relevant cloud service providers have been informed of the verified exposures to enable appropriate mitigation measures.

Ledger Customer Data Exposed After Global-e Payment Processor Cloud Incident

 

A fresh leak of customer details emerged, linked not to Ledger’s systems but to Global-e - an outside firm handling payments for Ledger.com. News broke when affected users received an alert email from Global-e. That message later appeared online, posted by ZachXBT, a known blockchain tracker using a fake name, via the platform X. 

Unexpectedly, a breach exposed some customer records belonging to Ledger, hosted within Global-e’s online storage system. Personal details, including names and email addresses made up the compromised data, one report confirmed. What remains unclear is the number of people impacted by this event. At no point has Global-e shared specifics about when the intrusion took place.  

Unexpected behavior triggered alerts at Global-e, prompting immediate steps to secure systems while probes began. Investigation followed swiftly after safeguards were applied, verifying unauthorized entry had occurred. Outside experts joined later to examine how the breach unfolded and assess potential data exposure. Findings showed certain personal details - names among them - were viewed without permission. Contact records also appeared in the set of compromised material. What emerged from analysis pointed clearly to limited but sensitive information being reached. 

Following an event involving customer data, Ledger confirmed details in a statement provided to CoinDesk. The issue originated not in Ledger's infrastructure but inside Global-e’s operational environment. Because Global-e functions as the Merchant of Record for certain transactions, it holds responsibility for managing related personal data. That role explains why Global-e sent alerts directly to impacted individuals. Information exposed includes records tied to purchases made on Ledger.com when buyers used Global-e’s payment handling system. 

While limited to specific order-related fields, access was unauthorized and stemmed from weaknesses at Global-e. Though separate entities, their integration during checkout links them in how transactional information flows. Customers involved completed orders between defined dates under these service conditions. Security updates followed after discovery, coordinated across both organizations. Notification timing depended on forensic review completion by third-party experts. Each step aimed at clarity without premature disclosure before full analysis. 

Still, the firm pointed out its own infrastructure - platform, hardware, software - was untouched by the incident. Security around those systems remains intact, according to their statement. What's more, since users keep control of their wallets directly, third parties like Global-e cannot reach seed phrases or asset details. Access to such private keys never existed for external entities. Payment records, meanwhile, stayed outside the scope of what appeared in the leak. 

Few details emerged at first, yet Ledger confirmed working alongside Global-e to deliver clear information to those involved. That setup used by several retailers turned out to be vulnerable, pointing beyond a single company. Updates began flowing after detection, though the impact spread wider than expected across shared infrastructure. 

Coming to light now, this revelation follows earlier security problems connected to Ledger. Back in 2020, a flaw at Shopify - the online store platform they used - led to a leak affecting 270,000 customers’ details. Then, in 2023, another event hit, causing financial damage close to half a million dollars and touching multiple DeFi platforms. Though different in both scale and source, the newest issue highlights how reliance on outside vendors can still pose serious threats when handling purchases and private user information.  

Still, Ledger’s online platforms showed no signs of a live breach on their end, yet warnings about vigilance persist. Though nothing points to internal failures, alerts remind customers to stay alert regardless. Even now, with silence across official posts, guidance leans toward caution just the same.

Neo AI Browser: How Norton’s AI-Driven Browser Aims to Change Everyday Web Use

 


Web browsers are increasingly evolving beyond basic internet access, and artificial intelligence is becoming a central part of that shift. Neo, an AI-powered browser developed by Norton, is designed to combine browsing, productivity tools, and security features within a single platform. The browser positions itself as a solution for users seeking efficiency, privacy control, and reduced online distractions.

Unlike traditional browsers that rely heavily on cloud-based data processing, Neo stores user information directly on the device. This includes browsing history, AI interactions, and saved preferences. By keeping this data local, the browser allows users to decide what information is retained, synchronized, or removed, addressing growing concerns around data exposure and third-party access.

Security is another core component of Neo’s design. The browser integrates threat protection technologies intended to identify and block phishing attempts, malicious websites, and other common online risks. These measures aim to provide a safer browsing environment, particularly for users who frequently navigate unfamiliar or high-risk websites.

Neo’s artificial intelligence features are embedded directly into the browsing experience. Users can highlight text on a webpage to receive simplified explanations or short summaries, which may help when reading technical, lengthy, or complex content. The browser also includes writing assistance tools that offer real-time grammar corrections and clarity suggestions, supporting everyday tasks such as emails, reports, and online forms.

Beyond text-based tools, Neo includes AI-assisted document handling and image-related features. These functions are designed to support content creation and basic processing tasks without requiring additional software. By consolidating these tools within the browser, Neo aims to reduce the need to switch between multiple applications during routine work.

To improve usability, Neo features a built-in ad blocker that limits intrusive advertising. Reducing ads not only minimizes visual distractions but can also improve page loading speeds. This approach aims to provide a smoother and more focused browsing experience for both professional and casual use.

Tab management is another area where Neo applies automation. Open tabs are grouped based on content type, helping users manage multiple webpages more efficiently. The browser also remembers frequently visited sites and ongoing tasks, allowing users to resume activity without manually reorganizing their workspace.

Customization plays a role in Neo’s appeal. Users can adjust the browser’s appearance, create shortcuts, and modify settings to better match their workflow. Neo also supports integration with external applications, enabling notifications and tool access without leaving the browser interface.

Overall, Neo reflects a broader trend toward AI-assisted browsing paired with stronger privacy controls. By combining local data storage, built-in security, productivity-focused AI tools, and performance optimization features, the browser presents an alternative approach to how users interact with the web. Whether it reshapes mainstream browsing habits remains to be seen, but it underlines how AI is steadily redefining everyday digital experiences.



Hyundai Faces Security Incident With Potential Data Exposure

 


In the past few months, Hyundai AutoEver America, a division of Hyundai Motor Group, has confirmed a recent data breach that exposed sensitive personal information after hackers infiltrated its internal IT environment earlier this year, revealing a recent data breach. 

A company spokesperson told me that unauthorized access to the company's computer systems began on February 22, 2025 and went undetected until March 2, giving intruders nine days to access confidential data. 

The early breach notices didn't specify how many people were affected, but according to state regulatory disclosures as well as a subsequent statement issued to Kelley Blue Book, approximately 2,000 people—out of the over 2.7 million users HAEA serves across Hyundai, Kia, and Genesis platforms—were impacted. There have been a number of compromises of the data, including names, Social Security numbers, and driving license information. 

In response to the suspicious activity, HAEA contacted an external cybersecurity expert who conducted an investigation, contained the intrusion, and informed law enforcement. As officials continue to assess the full scope of the incident, officials have begun issuing formal notices to those whose information was possibly exposed. 

It was only in the months that followed that it became increasingly clearer and more troubling just what the breach's consequences and the broader risks associated with connected vehicles were in the future. Even though Hyundai AutoEver America eventually acknowledged that the incident could have affected as many as 2.7 million Hyundai, Kia, and Genesis owners, internal assessments and state filings later narrowed the directly affected group to merely 2,000 individuals, yet the sensitive nature of the data involved makes even this smaller number quite significant. 

A nine-day intrusion that took place between February 22 and March 2, 2025, revealed the names, addresses, phone numbers, driver’s license numbers, and Social Security numbers of several automobile manufacturers, revealing to intruders a full range of data and details that underpinned core digital services across the automaker’s brands during that period. 

Among privacy experts, there is no doubt that what has caused concern is not just the scope of information but also that it has taken seven months for customers to be informed about the incident, a timeframe that gave the possibility for stolen identities to be misused or combined with other data circulating from other breaches.

Hyundai is also experiencing a growing pattern of security breaches since 2023, which reinforces concerns that these are not isolated incidents but rather signs of deeper structural problems. As the episode illustrates, modern cars—once purely mechanical devices—now act as sophisticated data hubs, collecting everything from passengers’ financial details to route histories, biometric inputs, driving behaviour, and even information synced from their mobile devices, which is not visible to the driver. 

Manufacturers are expanding their digital ecosystems and the breach has raised questions about the industry's ability to safeguard the vast and intimate data it collects on a regular basis. Immediately following the intrusion, Hyundai AutoEver America made an effort to reassure its customers by offering two years of complimentary identity theft and credit monitoring services through Epiq as a gesture of goodwill.

In spite of this, security analysts note that such measures are rarely sufficient to relieve customers after sensitive information has been stolen. Additionally, Hyundai Motor Europe’s disclosure also brought back memories of a similar experience it suffered just a year earlier when it was attacked by a ransomware gang called Black Basta, which claimed to have taken over 3TB of internal files before appearing dormant in early 2025, when the company lost control of its operations. 

All in all, these incidents emphasize one more uncomfortable reality: automakers now harvest and manage far greater amounts of personal information than most drivers are aware of. Besides the information required for financing or registration of vehicles, companies routinely collect (and in some cases monetize) data regarding the locations of their customers, their driving habits, the biometric patterns they use, and even behavioral patterns that can help them infer consumers' preferences with a remarkable degree of accuracy. 

Following a complaint made by General Motors that it had shared driver data with third-parties to the point of being able to obtain their information from them, the Federal Trade Commission issued a five-year ban on the practice. In July, a U.S. Senate inquiry raised concerns about other manufacturers continuing the same data-sharing practices. 

The HAEA notified the California Attorney General of the incident by notifying them that they had enlisted cybersecurity experts to determine the scope of the breach and confirm that the intrusion had been contained, even though investigators were unable to determine if the information was exfiltrated. Those affected customers have been given 90 days to enroll in monitoring services, and a hotline has also been established to assist customers. 

As Hyundai AutoEver asserts, only a small number of users have been directly impacted by this incident, but the incident has ignited a wider industry debate over precisely how well automakers secure the ever-increasing amount of personal data embedded in most connected vehicles today. After Hyundai AutoEver America found out that a wide range of sensitive data points had been exposed as part of this breach, including a number of customer names, government-issued identification numbers, and passwords, it confirmed that the investigation of the technical footprint was continuing. 

Among the records that were compromised, according to notification letters sent to the individuals affected, were Social Security numbers and driver's license information, with each recipient receiving a customized breakdown of which data elements applied to them in the initial notification. In order to conduct the analysis in a comprehensive way, extensive forensic work and collaboration with external cybersecurity specialists were necessary. 

These specialists helped Hyundai AutoEver reconstruct the intrusion, assess database exposure, and determine which users needed formal notification. Hyundai AutoEver said it immediately terminated the intruder's access and implemented additional safeguards and was continuing to implement a comprehensive remediation program that was intended to prevent similar incidents in the future. 

Consequently, Epiq Privacy Solutions has been contacted by the company to offer complimentary two-year credit monitoring and identity protection services to impacted customers, which will include three-bureau monitoring and fraud detection tools, as well as a 90-day enrollment period. It should be noted that these protections are only a layer of protection, however, according to security experts. 

As a precautionary measure, they advise their customers to review financial statements, to check their credit reports, and to place fraud alerts or credit freezes with the major credit bureaus to reduce the risk of unauthorized account openings. 

In addition, this incident has brought about renewed discussions about digital hygiene for vehicle owners, ranging from updating passwords and enabling multifactor authentication on connected car applications to avoiding stored payment information in the infotainment system.

There are a number of cybercrime analysts who note that incidents of this nature often open the door to secondary scams, as cybercriminals impersonate automakers' support teams in order to steal more personal information from car owners through pages pretending to be account verifications and security updates. 

These developments have been identified by industry observers as part of a dramatic shift in the way in which cars now collect far more information than most drivers are aware of. These include location histories, biometric identifiers, behavioral patterns, and synced mobile data, to name a few. 

The results of this study indicate that consumers should adopt strong cybersecurity practices, including using reputable antivirus software, staying current on device updates, and thinking about data-removal solutions that will reduce exposure to data-broker websites as a result of data misuse. Several automakers have been affected by this new trend; the Federal Trade Commission imposed a five-year ban on General Motors' ability to sell data on drivers earlier this year. 

Additionally, a Senate investigation has raised concerns about similar practices in other automakers, including Hyundai, as well. In spite of Hyundai AutoEver's assertion that only a relatively small number of its customers were directly affected by this breach, the incident has brought to light questions about the effectiveness with which carmakers are safeguarding the growing amounts of data embedded in connected cars, as well as what consumers should do in the rapidly growing digital world in order to protect themselves from the threat of fraud. 

It is clear from the Hyundai AutoEver breach that the automobile industry needs to rethink how it approaches data security in an increasingly interconnected digital age, where vehicles become increasingly interconnected digital ecosystems. It is important to note that meaningful protection depends both on stronger corporate safeguards as well as on proactive vigilance on the part of drivers in light of increased regulatory oversight and consumers' increasing awareness of how their information is being used.

It is vital that consumers play an important role in reducing future risks by practicing stricter digital hygiene, minimizing unnecessary data sharing, and demanding that automakers communicate their information more clearly, in order to ensure that the convenience of connected cars does not come at the expense of their individual privacy rights.

Indian Tax Department Fixes Major Security Flaw That Exposed Sensitive Taxpayer Data

 

The Indian government has patched a critical vulnerability in its income tax e-filing portal that had been exposing sensitive taxpayer data to unauthorized users. The flaw, discovered by security researchers Akshay CS and “Viral” in September, allowed logged-in users to access personal and financial details of other taxpayers simply by manipulating network requests. The issue has since been resolved, the researchers confirmed to TechCrunch, which first reported the incident. 

According to the report, the vulnerability exposed a wide range of sensitive data, including taxpayers’ full names, home addresses, email IDs, dates of birth, phone numbers, and even bank account details. It also revealed Aadhaar numbers, a unique government-issued identifier used for identity verification and accessing public services. TechCrunch verified the issue by granting permission for the researchers to look up a test account before confirming the flaw’s resolution on October 2. 

The vulnerability stemmed from an insecure direct object reference (IDOR) — a common but serious web flaw where back-end systems fail to verify user permissions before granting data access. In this case, users could retrieve another taxpayer’s data by simply replacing their Permanent Account Number (PAN) with another PAN in the network request. This could be executed using simple, publicly available tools such as Postman or a browser’s developer console. 

“This is an extremely low-hanging thing, but one that has a very severe consequence,” the researchers told TechCrunch. They further noted that the flaw was not limited to individual taxpayers but also exposed financial data belonging to registered companies. Even those who had not yet filed their returns this year were vulnerable, as their information could still be accessed through the same exploit. 

Following the discovery, the researchers immediately alerted India’s Computer Emergency Response Team (CERT-In), which acknowledged the issue and confirmed that the Income Tax Department was working to fix it. The flaw was officially patched in early October. However, officials have not disclosed how long the vulnerability had existed or whether it had been exploited by malicious actors before discovery. 

The Ministry of Finance and the Income Tax Department did not respond to multiple requests for comment on the breach’s potential scope. According to public data available on the tax portal, over 135 million users are registered, with more than 76 million having filed returns in the financial year 2024–25. While the fix has been implemented, the incident highlights the critical importance of secure coding practices and stronger access validation mechanisms in government-run digital platforms, where the sensitivity of stored data demands the highest level of protection.

Telstra Denies Scattered Spider Data Breach Claims Amid Ransom Threats

 

Telstra, one of Australia’s leading telecommunications companies, has denied claims made by the hacker group Scattered Spider that it suffered a massive data breach compromising nearly 19 million personal records. The company issued a statement clarifying that its internal systems remain secure and that the data in question was scraped from publicly available sources rather than stolen. In a post on X (formerly Twitter), Telstra emphasized that no passwords, banking details, or sensitive identification data such as driver’s licenses or Medicare numbers were included in the dataset. 

The claims originated from a dark web post published on October 3 by a group calling itself Scattered Lapsus$ Hunters, an offshoot of Scattered Spider. The group alleged it had stolen more than 100GB of personally identifiable information, including names and physical addresses, and warned that company executives should negotiate to avoid further data exposure. The attackers claimed the alleged breach took place in July 2023 and threatened to release the data publicly if a ransom was not paid by October 13, 2025. They also asserted possession of over 16 million records contained in a file named telstra.sql, which they said was part of a larger collection of 19 million records. 

In a surprising twist, the ransom note also mentioned Salesforce, the global cloud computing company, demanding negotiations begin with its executives. Salesforce swiftly rejected the demand, issuing a statement on October 8 declaring that it “will not engage, negotiate with, or pay any extortion demand,” aligning with global cybersecurity guidelines that discourage ransom payments. 

Scattered Lapsus$ Hunters has made similar claims about breaches involving several major corporations, including Qantas, IKEA, and Google AdSense. Cybersecurity intelligence platforms like Cyble Vision have documented multiple previous instances of alleged Telstra data breaches, some dating back to 2022. In one notable case, a threat actor called UnicornLover67 claimed to possess a dataset containing over 47,000 Telstra employee records, including email addresses and hashed passwords. Telstra has previously confirmed smaller breaches linked to third-party service providers, most recently in 2022, affecting around 132,000 customers. 

However, cybersecurity analysts remain uncertain whether the current claims represent a fresh breach or a recycling of old data. Experts suggest that previously leaked or publicly available datasets may have been repurposed to appear as new evidence of compromise. This possibility aligns with Telstra’s statement that no recent intrusion has occurred. 

The investigation into the alleged breach remains ongoing as the ransom deadline approaches. While Telstra continues to assert that its systems are uncompromised, the persistence of repeated breach claims underscores the growing challenge of misinformation and data reuse in the cybercrime landscape. The Cyber Express has reached out to Telstra for further updates and will continue to monitor the situation as new details emerge.

Sam Altman Pushes for Legal Privacy Protections for ChatGPT Conversations

 

Sam Altman, CEO of OpenAI, has reiterated his call for legal privacy protections for ChatGPT conversations, arguing they should be treated with the same confidentiality as discussions with doctors or lawyers. “If you talk to a doctor about your medical history or a lawyer about a legal situation, that information is privileged,” Altman said. “We believe that the same level of protection needs to apply to conversations with AI.”  

Currently, no such legal safeguards exist for chatbot users. In a July interview, Altman warned that courts could compel OpenAI to hand over private chat data, noting that a federal court has already ordered the company to preserve all ChatGPT logs, including deleted ones. This ruling has raised concerns about user trust and OpenAI’s exposure to legal risks. 

Experts are divided on whether Altman’s vision could become reality. Peter Swire, a privacy and cybersecurity law professor at Georgia Tech, explained that while companies seek liability protection, advocates want access to data for accountability. He noted that full privacy privileges for AI may only apply in “limited circumstances,” such as when chatbots explicitly act as doctors or lawyers. 

Mayu Tobin-Miyaji, a law fellow at the Electronic Privacy Information Center, echoed that view, suggesting that protections might be extended to vetted AI systems operating under licensed professionals. However, she warned that today’s general-purpose chatbots are unlikely to receive such privileges soon. Mental health experts, meanwhile, are urging lawmakers to ban AI systems from misrepresenting themselves as therapists and to require clear disclosure when users are interacting with bots.  

Privacy advocates argue that transparency, not secrecy, should guide AI policy. Tobin-Miyaji emphasized the need for public awareness of how user data is collected, stored, and shared. She cautioned that confidentiality alone will not address the broader safety and accountability issues tied to generative AI. 

Concerns about data misuse are already affecting user behavior. After a May court order requiring OpenAI to retain ChatGPT logs indefinitely, many users voiced privacy fears online. Reddit discussions reflected growing unease, with some advising others to “assume everything you post online is public.” While most ChatGPT conversations currently center on writing or practical queries, OpenAI’s research shows an increase in emotionally sensitive exchanges. 

Without formal legal protections, users may hesitate to share private details, undermining the trust Altman views as essential to AI’s future. As the debate over AI confidentiality continues, OpenAI’s push for privacy may determine how freely people engage with chatbots in the years to come.

Millions of Resumes Exposed Online Due to Unsecured Hiring Platform

 



A major data exposure has come to light after cybersecurity experts discovered an unsecured online storage system containing nearly 26 million documents, many of which appear to be resumes of job seekers in the United States.

The exposed files were found in a cloud-based storage system, commonly used to save and share digital files. According to the research team, this storage space had not been properly secured, meaning anyone who knew where to look could access its contents without needing a password or any special permissions.

On further examination, it was revealed that the majority of the documents stored in the system were personal resumes and CVs. These files included sensitive personal details like full names, phone numbers, email addresses, education history, previous work experience, and other professional information. In the wrong hands, such detailed personal data can become a serious security risk.

Experts warn that job seekers are particularly vulnerable in situations like this. If cybercriminals gain access to such data, they can use it to send highly personalized scam messages. These messages may appear trustworthy, as they can be tailored using real employment history or job interests, making it easier to trick someone into clicking a malicious link or sharing their login information.

One common tactic includes sending fake job offers or interview invitations that secretly install harmful software on a person’s device. Some advanced scams may even go as far as conducting fake job interviews before sending victims "sample tasks" that involve downloading malware.

The database in question was linked to a platform used by employers and hiring teams to manage job applications and connect with candidates. However, the researchers who found the issue say they did not receive any confirmation that access to the exposed files has been blocked. While the team reached out to suggest tightening security settings, it’s unclear whether any action was taken.

There is no current proof that the data has been used by cybercriminals yet, but experts note that the longer the files remain unprotected, the higher the risk of misuse. Even if no signs of abuse have appeared so far, the availability of such information online creates an ongoing threat.

This situation serves as a reminder for companies handling sensitive data to prioritize cybersecurity. Properly configuring cloud storage, regularly updating access settings, and limiting who can view certain files are essential steps in preventing such exposures. It’s not just about protecting a system, it’s about safeguarding real people’s identities and futures.