Kaspersky has reported a spike in phishing emails containing malicious QR codes. Detections for these jumped from 46,969 in August 2025 to 249,723 in November 2025 – a more than fivefold growth – as cybercriminals increasingly exploit QR codes, a trend that will likely continue in 2026.

Attackers use QR codes in emails more frequently because they provide a simple and cost-effective way to conceal malicious URLs, evading detection by many protective solutions.

These QR codes are often embedded directly in email bodies or, even more commonly, within PDF attachments – an evolution that both masks phishing links and encourages users to scan them on mobile phones, which may have weaker security than work PCs.

Malicious QR codes commonly appear in mass phishing campaigns as well as targeted ones. Links embedded within them may lead to:

• Phishing forms impersonating login pages for services like Microsoft accounts or internal corporate portals, designed to steal usernames, passwords, and other credentials.
• Fake HR notifications urging employees to review or sign documents, such as vacation schedules, or even view lists of terminated staff, ultimately directing to credential-stealing sites.
• Fraudulent invoices or purchase confirmations in PDF attachments, often combined with vishing (voice phishing) tactics that prompt victims to call provided phone numbers to “cancel” or clarify the transaction, enabling further social engineering attacks.

These tactics exploit trust in routine business communications, leading to credential theft, account takeovers, data breaches, and financial fraud.

“Malicious QR codes have evolved into one of the most effective phishing tools, particularly when hidden in PDF attachments or disguised as legitimate business communications like HR updates.

“The explosive growth in November 2025 highlights how attackers are capitalising on this low-cost evasion technique to target employees on mobile devices, where protection is often minimal.

“Without advanced image analysis at the email gateway and safe scanning practices, organisations are left vulnerable to credential compromise and downstream breaches,” comments Roman Dedenok, Anti-Spam Expert at Kaspersky.

To defend against this escalating threat, Kaspersky recommends educating employees on cybersecurity and deploying a mail server security solution such as Kaspersky Security for Mail Server that provides trusted and secure corporate email exchange, countering spam, email-borne infections, all forms of phishing, business email compromise (BEC), QR code attacks, and other threats.

Kaspersky has detected a scam tactic leveraging the OpenAI platform. Attackers are abusing OpenAI’s organisation creation and team invitation features to send spam emails from legitimate OpenAI addresses, potentially tricking users into clicking scam links or calling fraudulent phone numbers.

The spam campaign begins with attackers registering an account on the OpenAI platform. During registration, users are prompted to enter an organisation name, which can consist of any combination of symbols. Scammers exploit this by embedding deceptive text and fraudulent links or phone numbers directly into the field for organisation name itself.

Once the “organisation” is created, OpenAI provides an option to “invite your team,” allowing the input of target email addresses of victims. When invitations are sent, they originate from OpenAI’s address, making them appear fully legitimate from a technical standpoint.

Kaspersky detected several types of messages containing email threats sent in such a way. These are scam emails that promote fraudulent offers, such as adult services. Another attack angle is vishing – false notifications claiming a subscription has been renewed for a large sum: attackers instruct recipients to call a provided phone number to “cancel” the charge or take other actions that lead to further compromise. There may also be other email threats spreading via OpenAI platform.

The text that the attackers want the victims to read (highlighted in bold in the email template) is structurally inconsistent with the rest of the email template – which was originally designed to invite project collaborators. But the attackers bet on the fact that the victims would not pay attention.

“This case highlights a vulnerability in how platform features can be weaponised for social engineering email attacks. By embedding deceptive elements in seemingly innocuous fields like organisation names, scammers attempt to bypass traditional email filters and exploit user trust in reputable services.

“We urge all users to verify invitations carefully and avoid clicking embedded links without scrutiny. We also recommend brands to consider whether their online services or platforms could be abused by attackers,” comments Anna Lazaricheva, senior spam analyst at Kaspersky.

In 2025, the retail and e-commerce sector continued to face intense pressure from cybercriminals. According to Kaspersky data, 14,41%* of users in the global retail sector encountered web-based threats, while 22,20% were affected by on-device attacks.

Ransomware remains a serious concern for the industry. Last year, 8,25% of retail and e-commerce companies experienced ransomware incidents, and the number of unique B2B users in the sector affected by ransomware detections rose by 152% compared to 2023, signalling a sharp escalation in targeted attacks.

Phishing also continues to be a major threat vector. Kaspersky identified 6.7 million phishing attacks targeting users of online stores, delivery services, and payment systems in 2025. More than half of these attacks (50,58%) were aimed specifically at online stores, underscoring cybercriminals’ focus on e-commerce platforms as high-value targets for fraud and data theft.

A look at 2025 cybersecurity for retail & e-commerce: Trends and what happened

A stealer with a taste for pizza delivery. Shopping and food ordering via mobile apps are routine user behaviours. However, 2025 demonstrated that even downloading a seemingly legitimate app from an official app store does not guarantee safety, nor does it ensure that user data and financial credentials will not be compromised.

Ransomware detections in the B2B sector increased due to a single dominant actor. The number of unique users in the Retail & E-commerce sector who encountered ransomware detections increased by 152% in 2025 compared to 2023 (Nov 2024 – Oct 2025 vs. Nov 2022 – Oct 2023).

The most significant growth occurred during the 2024-2025 period and is largely attributable to the rapid spread of the Trojan-Ransom.Win32. Dcryptor family, which became highly prevalent across the retail and e-commerce sector in some of the analysed markets. This malware is a trojanised ransomware variant that leverages the legitimate DiskCryptor utility to encrypt disk partitions on victim systems.

Phishing activity in the online retail segment stood out. Despite being a long-established attack technique, phishing remains highly prevalent in the context of online purchasing.

From November 2024 through to October 2025, Kaspersky products blocked 6,651,955 attempts to access phishing links targeting users of online stores, payment systems, and delivery services. Of these attempts, 50.58% targeted online shoppers, 27.3% impersonated payment systems, and 22.12% targeted users of delivery companies.

Sales seasons continue to do the work for attackers. Seasonal peaks in online shopping consistently provide attackers with predictable opportunities to scale user-focused attacks.

Periods of heightened promotional activity lower user vigilance and allow familiar phishing and spam scenarios to blend into legitimate marketing traffic, increasing their overall effectiveness.

Predictions: What retail & e-commerce cybersecurity might face in 2026

Chatbots are likely to become a common product discovery tool across online marketplaces. Unlike traditional search, conversational interfaces encourage users to share more detailed, natural-language requests, revealing preferences, constraints, and contextual information.

This shift expands the privacy attack surface, as platforms accumulate richer user profiles through chat interactions. As a result, chatbot logs may become as sensitive as transactional data, increasing the risks of over-collection, misuse, or exposure of personal information.

“Search itself is changing, including how people look for products online. In 2025, there was a gradual shift from simple keyword queries to more conversational and visual ways of finding what to buy. As these models rely on broader user input, careful handling of the data involved will remain an important consideration for maintaining user trust,” comments Anna Larkina, Web data and privacy analysis expert at Kaspersky.

Changes in taxes and trade rules might be exploited in online fraud. Modifications in taxes, import duties, and cross-border trade rules are likely to be used as lures in phishing campaigns and fraudulent online stores, promoting unrealistically cheap offers or claims of avoided fees.

As pricing and fee rules continue to evolve across markets, it may lower vigilance, increasing the effectiveness of such schemes, particularly against small and mid-sized retailers.

AI-powered shopping assistants are expected to increasingly operate outside retail platforms, embedding themselves into browsers, mobile apps, and third-party services. While designed to simplify navigation and price discovery, these tools shift data collection beyond the retailer’s perimeter, creating new and less visible privacy risks.

To function effectively, external AI shopping agents require continuous access to user behaviour, including browsing activity, search intent, location context and product interactions across multiple sites.

This enables the aggregation of detailed behavioural profiles outside the direct control of both users and retail platforms, increasing the risks of over-collection, opaque data usage, and unintended exposure.

Image-based product search might become a new challenge in privacy risks. Previously, the main privacy concern around user images in e-commerce was limited to photos voluntarily shared in product reviews.

However, image-based product search is expected to make photo uploads a routine part of the shopping experience across major retail platforms. While this feature improves product discovery, it also increases the risk of unintended exposure of personal data.

User-submitted images may contain faces, home environments, or sensitive details, such as names, phone numbers, or addresses visible on shipping labels or packaging, making secure processing, data minimisation, and limited retention critical requirements for retailers.