Kled AI, US-based developer, has announced the removal of its application from the Nigerian app store, alongside an IP restriction affecting the region, citing what it described as an “unmanageable level of fraudulent activity” on the platform.

Kled is a data marketplace that rewards users for uploading photos, videos, and other multimodal content.

Avi Patel, 22-year-old founder, in his X handle, said the decision followed months of internal review, during which the startup found that a large share of uploads from Nigeria, including images, documents, and videos meant for AI training, were fake, duplicated, or generated by artificial intelligence.

Kled operates what it describes as an opt-in data marketplace, where users voluntarily upload personal content in exchange for payment, with the material later sold to AI labs for training models.

The startup said it has paid hundreds of thousands of users globally and processed over one billion data assets within four months of launch.

However, Patel said Nigeria stood out negatively.

According to him, the company reviewed a sample of 10 million uploads from the country and found that only a small fraction met quality standards required for AI training.

He added that the problem escalated when the platform was flooded with manipulated identity documents, including fake passports, during its verification process.

“As a startup, we cannot absorb the cost of filtering that level of bad data,” Patel said, noting that the company has now removed the app from Nigeria’s Apple App Store and imposed an IP ban on the region while it strengthens its fraud detection systems.

“On top of all of this, every time we make a post there is someone asking us to bring the region back within seconds. We hear you, but it’s gotten out of hand,” he added.

Despite the suspension, the company maintained that the move is temporary and not permanent.

“We’ve made this decision with great care. We love everyone who has genuinely supported Kled from Nigeria, and we hope to return when the time is right,” the statement concluded.

The decision has triggered backlash among Nigerian users, many of whom accuse the company of stereotyping and unfairly targeting the country.

Patel, however, insists the move is purely business-driven and not linked to race or nationality, stressing that Kled remains available in other African markets.

.

Although the main initial vectors in 2025 remain similar to 2024, their combined share has grown to over 80%. Public-facing applications account for 43.7%, while trusted relationships have increased from 12.7% to 15.5%.

Valid accounts make up 25.4%. These insights are from the recent Global Report by Kaspersky Security Services.

The ‘Anatomy of a Cyber World’ is an in-depth global report based on incident data gathered in 2025 from Kaspersky Managed Detection and Response, Kaspersky Incident Response, Kaspersky Compromise Assessment and Kaspersky SOC Consulting.

It highlights the most common attacker tactics, techniques and tools, as well as the peculiarities of detected incidents and their distribution across regions and industries.

According to data derived from Kaspersky Incident Response, the top three initial attack vectors have remained relatively stable over the past seven years and have not changed significantly. Valid accounts and exploits in public-facing applications consistently represent the most common entry points.

The third position has periodically shifted: malicious emails, once a common initial vector, were replaced by trusted relationships, which first appeared in 2021 and entered the TOP-3 in 2023. By 2025, the distribution of main vectors looked as follows:

These attack vectors are often interconnected within the same chain, for example, organisations compromised through trusted relationships are frequently first breached via exploits in public-facing applications. Recent cases reveal attackers targeting service providers or IT integrators to then access their clients.

This problem is compounded by many small service providers lacking dedicated cybersecurity expertise and resources. As they manage accounting software or websites, breaches in these companies can lead to the compromise of their clients’ systems through exploited remote access.

When examining the investigated attacks in terms of duration and impact, the data shows that the majority (50.9%) of them were rapid in nature, typically lasting less than a day and most often resulting in file encryption.

A significant portion (33%) were long-lasting, with an average duration of 108 hours, during which attackers not only encrypted files but also installed persistence mechanisms, compromised Active Directory and caused data leakage.

The remaining 16.1% exhibited a hybrid pattern: they initially appeared as rapid attacks but involved a considerable delay between the initial breach and subsequent malicious activities, extending their overall duration to nearly 19 days.

“Given that attackers are increasingly orchestrating coordinated, multi-stage attacks, organisations cannot afford to rely on a reactive, “firefighting” approach. To counter this, a proactive security posture is essential, one that embeds real-time threat monitoring and continuous detection into everyday operations.

This enables defenders to respond swiftly to adversary activity before it escalates. Key measures for protecting digital assets against both rapid intrusions and long-term compromises include: timely patching, enforcement of multi-factor authentication and strict control of third-party access,” comments Konstantin Sapronov, Head of Global Emergency Response Team at Kaspersky.

Meta has announced new advancements in its age assurance technology as part of its ongoing efforts to create safer, age-appropriate experiences for young people across its platforms. Through a combination of AI, product design, and parental support tools, Meta continues to strengthen how it identifies teens, protects them by default, and supports families in navigating digital environments.

Meta

Strengthening underage enforcement with advanced AI

Meta requires users to be at least 13 years old to use its platforms and continues to invest in advanced technologies to uphold this policy at scale. As part of these efforts, the company is further enhancing its AI-driven systems to more effectively identify and take action on accounts that may belong to underage users.

These advancements include:

• Contextual AI analysis across profiles: Meta’s systems analyse a wide range of signals—including posts, comments, bios and captions—to identify contextual indicators such as references to school environments or age-related milestones. This capability is being expanded across additional surfaces within Meta’s apps, strengthening enforcement in a more consistent and proactive way.

• Advanced visual analysis technology: Meta is introducing AI that can interpret general age-related cues within photos and videos. This technology estimates age ranges based on broad characteristics and does not use facial recognition or identify individuals. When combined with behavioural and textual signals, it significantly enhances detection accuracy.

• Expanded enforcement and verification processes: Accounts identified as potentially underage are subject to age verification requirements. Where age cannot be confirmed, accounts may be removed to maintain platform integrity.

• Improved reporting and flagging tools: Meta is making it easier for people to report suspected underage accounts through simplified reporting flows available both in-app and via the Help Center, helping surface potential violations more efficiently.

• AI-supported review systems: To improve consistency and speed, Meta is supplementing human review teams with AI models that apply standardised evaluation criteria to reports, enabling faster and more reliable enforcement outcomes.

• Stronger circumvention safeguards: Meta is also enhancing its ability to detect and prevent repeat attempts by users who may try to bypass age restrictions by creating new accounts.

While many of these AI-driven systems are already in use globally, certain advanced capabilities continue to be rolled out progressively across additional markets.

Expanding Teen Account protections

Meta continues to expand its Teen Account framework, which is designed to provide built-in protections that limit unwanted contact and reduce exposure to inappropriate content. Since its introduction, hundreds of millions of teens have been enrolled in these protections across Instagram, Facebook, and Messenger.

These protections include automatically placing teens under 18 into age-appropriate experiences, including a default 13+ content setting designed to limit exposure to sensitive content.

Building on this progress, Meta is further scaling its proactive detection technology that identifies users who may be teens—even if they have entered an adult birthdate—and automatically places them into age-appropriate settings. This technology, already rolled out in several markets, is being expanded to additional regions, with the goal of making these protections available more broadly over time.

Supporting parents with tools and guidance

Meta continues to support parents as key partners in helping teens navigate online experiences safely. The company is introducing new notifications and guidance designed to help parents better understand how to verify their teen’s age and encourage open conversations about the importance of providing accurate information online.

These efforts build on existing resources available through Meta’s Family Center, which provides tools and educational materials to help families manage their digital experiences more effectively.

Meta also maintains age verification requirements for users who attempt to change their age in ways that may bypass protections, using a combination of ID verification and facial age estimation tools.

Advocating for industry-wide solutions

Meta continues to emphasise that age assurance is a complex, industry-wide challenge that requires broader collaboration. The company supports approaches where age verification is conducted at the operating system or app store level, enabling developers to deliver consistent, age-appropriate experiences across apps.

In addition to AI-based detection, Meta uses age estimation based on user activity and signals, as well as user reports, to help determine whether someone may be misrepresenting their age.

Meta believes that such an approach would help reduce fragmentation, improve consistency in protections, and provide a more privacy-preserving solution compared to requiring each individual app to implement separate systems.