Network security Archives | Digital Watch Observatory

Deepfake attacks push organisations to rethink cybersecurity strategies

Organisations are strengthening their cybersecurity strategies as deepfake attacks become more convincing and easier to produce using generative AI.

Security experts alert that enterprises must move beyond basic detection tools and adopt layered security strategies to defend against the growing threat of deepfake attacks targeting communications and digital identity.

Many existing tools for identifying manipulated media are still imperfect. Digital forensics expert Hany Farid estimates that some systems used to detect deepfake attacks are only about 80 percent effective and often fail to explain how they determine whether an image, video, or audio recording is authentic. The lack of explainability also raises challenges for legal investigations and public verification of suspicious media.

Cybersecurity companies are creating new technologies to improve the detection of deepfake attacks by analysing slight signals that are difficult for humans to notice. Firms such as GetReal Security, Reality Defender, Deep Media, and Sensity AI examine lighting consistency, shadow angles, voice patterns, and facial movements. Environmental indicators such as device location, metadata, and IP information can also help security teams spot potential deepfake attacks.

However, experts say detection alone cannot fully protect organisations from deepfake attacks. Companies are increasingly conducting internal red-team exercises that simulate impersonation scenarios to expose weaknesses in verification procedures. Multi-factor authentication techniques can reduce the risk of employees responding to fraudulent communications.

Another emerging defence involves digital provenance systems designed to track the origin and modification history of digital content. Initiatives such as the Coalition for Content Provenance and Authenticity (C2PA) embed cryptographically signed metadata into media files, allowing organisations to verify whether content linked to suspected deepfake attacks has been altered.

Recent experiments highlight how testing these threats can be. In February, cybersecurity company Reality Defender conducted an exercise with NATO by introducing deepfake media into a simulated military scenario. The findings showed how easily even experienced officials can struggle to identify manipulated communications, reinforcing calls for automated systems capable to detecting deepfake attacks across critical infrastructure.

As generative AI tools continue to advance, organisations are expected to combine detection technologies, stronger verification procedures, and provenance tracking to reduce the risks posed by deepfake attacks.

Would you like to learn more about AI, tech, and digital diplomacy? If so, ask our Diplo chatbot!

Hackers target WhatsApp and Signal in global encrypted messaging attacks

Foreign state-backed hackers are targeting accounts on WhatsApp and Signal used by government officials, diplomats, military personnel, and other high-value individuals, according to a security alert issued by the Portuguese Security Intelligence Service (SIS).

Portuguese authorities described the activity as part of a global cyber-espionage campaign aimed at gaining access to sensitive communications and extracting privileged information from Portugal and allied countries. The advisory did not identify the origin of the suspected attackers.

The warning follows similar alerts from other European intelligence agencies. Earlier this week, Dutch authorities reported that hackers linked to Russia were conducting a global campaign targeting the messaging accounts of officials, military personnel, and journalists.

Security agencies say the attackers are not exploiting vulnerabilities in the messaging platforms themselves. Both WhatsApp and Signal rely on end-to-end encryption designed to protect the content of messages from interception.

Instead, the campaign focuses on social engineering tactics that trick users into granting access to their accounts. According to the SIS report, attackers use phishing messages, malicious links, fake technical support requests, QR-code lures, and impersonation of trusted contacts.

The agency also warned that AI tools are increasingly being used to make such attacks more convincing. AI can help impersonate support staff, mimic familiar voices or identities, and conduct more realistic conversations through messages, phone calls, or video.

Once attackers gain access to an account, they may be able to read private messages, group chats, and shared files via WhatsApp and Signal. They can also impersonate the compromised user to launch additional phishing attacks targeting the victim’s contacts.

The alert echoes a previous warning issued by the Cybersecurity and Infrastructure Security Agency (CISA), which reported that encrypted messaging apps are increasingly being used as entry points for spyware and phishing campaigns targeting high-value individuals.

Would you like to learn more about AI, tech, and digital diplomacy? If so, ask our Diplo chatbot!

Writer files lawsuit against Grammarly over AI feature using experts’ identities

A journalist has filed a class action lawsuit against Grammarly after the company introduced an AI feature that generated editorial feedback by imitating well-known writers and public figures without their permission.

The legal complaint was submitted by investigative journalist Julia Angwin, who argued that the tool unlawfully used the identities and reputations of authors and commentators.

The feature, known as ‘Expert Review’, produced automated critiques presented as if they came from figures such as Stephen King, Carl Sagan and technology journalist Kara Swisher.

Such a feature was available to subscribers paying an annual fee and was designed to simulate professional editorial guidance.

Critics quickly questioned both the quality of the generated feedback and the decision to associate the tool with real individuals who had not authorised the use of their names or expertise.

Technology writer Casey Newton tested the system by submitting one of his own articles and receiving automated feedback attributed to an AI version of Swisher. The response appeared generic, casting doubt on the value of linking such commentary to prominent personalities.

Following criticism from writers and researchers, the feature was disabled. Shishir Mehrotra, chief executive of Grammarly’s parent company Superhuman, issued a public apology while defending the broader concept behind the tool.

The lawsuit reflects growing tensions around AI systems that replicate creative styles or professional expertise.

As generative AI technologies expand across writing and publishing industries, questions surrounding consent, intellectual labour and identity rights are becoming increasingly prominent.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

UK watchdog demands stronger child safety on social platforms

The British communications regulator Ofcom has called on major technology companies to enforce stricter age controls and improve safety protections for children using online platforms.

The warning targets services widely used by young audiences, including Facebook, Instagram, Roblox, Snapchat, TikTok and YouTube.

Regulators said that despite existing minimum age policies, large numbers of children under the age of 13 continue to access platforms intended for older users.

According to Ofcom research, more than 70 percent of children aged 8 to 12 regularly use such services.

Authorities have asked companies to demonstrate how they will strengthen protections and ensure compliance with minimum age requirements.

Platforms must present their plans by 30 April, after which Ofcom will publish an assessment of their responses and determine whether further regulatory action is necessary.

The regulator also outlined several key areas requiring improvement.

Companies in the UK are expected to implement more effective age-verification systems, strengthen protections against online grooming and ensure that recommendation algorithms do not expose children to harmful content.

Another concern involves product development practices.

Ofcom warned that new digital features, including AI tools, should not be tested on children without adequate safety assessments. Platforms are required to evaluate potential risks before launching significant updates.

The measures are part of the UK’s broader regulatory framework introduced under the Online Safety Act, which aims to reduce exposure to harmful online material.

The law requires platforms to prevent children from accessing content linked to pornography, suicide, self-harm and eating disorders, while limiting the promotion of violent or abusive material in recommendation feeds.

Ofcom indicated that enforcement action may follow if companies fail to demonstrate meaningful improvements. Regulators argue that stronger safeguards are necessary to restore public trust and ensure that digital platforms prioritise child safety in their design and operation.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

EU competition regulators expand scrutiny across the entire AI ecosystem

Competition authorities in the EU are broadening their oversight of the AI sector, examining every layer of the technology’s value chain.

Speaking at a conference in Berlin, Teresa Ribera explained that regulators are analysing the full ‘AI stack’ instead of focusing solely on consumer applications.

According to the competition chief, scrutiny extends beyond visible AI tools to the systems that support them. Investigations are assessing underlying models, the data used to train those models, as well as cloud infrastructure and energy resources that power AI systems.

Regulatory attention has already reached the application layer.

The European Commission opened an investigation in 2025 involving Meta after concerns emerged that the company could restrict competing AI assistants on its messaging platform WhatsApp.

Following regulatory pressure, Meta proposed allowing rival AI chatbots on the platform in exchange for a fee. European regulators are now assessing the proposal to determine whether additional intervention is necessary to preserve fair competition in rapidly evolving digital markets.

Authorities have also examined concentration risks across other parts of the AI ecosystem, including the infrastructure layer dominated by companies such as Nvidia.

Regulators argue that effective competition oversight must address the entire technology stack as AI markets expand quickly.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

EU privacy watchdogs warn over US plans to expand traveller data collection

European privacy authorities have raised concerns about proposed changes to the Electronic System for Travel Authorisation that could require travellers to the US to disclose extensive personal information, including social media activity.

The European Data Protection Board, which coordinates national data protection authorities across the EU, sent a letter to the European Commission asking whether the institution plans to intervene or respond to the updated requirements.

A proposal that would apply to visitors entering the US through the visa-waiver programme for short stays of up to 90 days.

Under the proposed changes, travellers may be required to provide details about their social media accounts covering the previous five years.

Authorities could also request personal data about family members, including addresses, phone numbers and dates of birth, information that privacy regulators argue is unrelated to travel authorisation.

Watchdogs also questioned how EU citizens could exercise their data protection rights once such information is transferred to US authorities, particularly regarding storage periods and potential misuse.

Parallel negotiations between the EU and the US have also attracted attention.

Discussions around a potential Enhanced Border Security Partnerships framework could allow US authorities to seek access to biometric databases held by European countries, including facial scans and fingerprint records.

European privacy regulators warned that such measures could raise significant concerns regarding fundamental rights and personal data protection for travellers from the EU.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

BeatBanker malware targets Android users in Brazil

A new Android malware called BeatBanker is targeting users in Brazil through fake Starlink and government apps. The malware hijacks devices, steals banking credentials, tampers with cryptocurrency transactions, and secretly mines Monero.

Infection begins on phishing websites mimicking the Google Play Store or the ‘INSS Reembolso’ app. Users are tricked into installing trojanised APKs, which evade detection through memory-based decryption and by blocking analysis environments.

Fake update screens maintain persistence while silently downloading additional malicious payloads.

BeatBanker initially combined a banking trojan with a cryptocurrency miner. It uses accessibility permissions to monitor browsers and crypto apps, overlaying fake screens to redirect Tether and other crypto transfers.

A foreground service plays silent audio loops to prevent the device from shutting down, while Firebase Cloud Messaging enables remote control of infected devices.

The latest variant replaces the banking module with the BTMOB RAT, providing full control over devices. Capabilities include automatic permissions, background persistence, keylogging, GPS tracking, camera access, and screen-lock credential capture.

Kaspersky warns that BeatBanker demonstrates the growing sophistication of mobile threats and multi-layered malware campaigns.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

AI browsers expose new cybersecurity attack surfaces

Security researchers have demonstrated that agentic browsers, powered by AI, may introduce new cybersecurity vulnerabilities.

Experiments targeting the Comet AI browser, developed by Perplexity AI, showed that attackers could manipulate the system into executing phishing scams in only a few minutes.

The attack exploits the reasoning process used by AI agents when interacting with websites. These systems continuously explain their actions and observations, revealing internal signals that attackers can analyse to refine malicious strategies and bypass built-in safeguards.

Researchers showed that phishing pages can be iteratively trained using adversarial machine learning methods, such as Generative Adversarial Networks.

By observing how the AI browser responds to suspicious signals, attackers can optimise fraudulent pages until the system accepts them as legitimate.

The findings highlight a shift in the cybersecurity threat landscape. Instead of deceiving human users directly, attackers increasingly focus on manipulating the AI agents that perform online actions on behalf of users.

Security experts warn that prompt injection vulnerabilities remain a fundamental challenge for large language models and agentic systems.

Although new defensive techniques are being developed, researchers believe such weaknesses may remain difficult to eliminate.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

AI agents face growing prompt injection risks

AI developers are working on new defences against prompt-injection attacks that aim to manipulate AI agents. Security specialists warn that attackers are increasingly using social engineering techniques to influence AI systems that interact with online content.

Researchers say AI agents that browse the web or handle user tasks face growing risks from hidden instructions embedded in emails or websites. Experts in the US note that attackers often attempt to trick AI into revealing sensitive information.

Engineers are responding by designing systems that limit the impact of manipulation attempts. Developers in the US say AI tools must include safeguards preventing sensitive data from being transmitted without user approval.

Security teams are also introducing technologies that detect risky actions and prompt users for confirmation. Specialists argue that strong system design and user oversight will remain essential as AI agents gain more autonomy.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

Google outlines roadmap for safer generative AI for young users

Google has presented a strategy for developing generative AI systems designed to protect younger users better better while supporting learning and creativity.

The approach emphasises building conversational AI experiences that balance innovation with safeguards tailored to children and teenagers.

The company’s framework rests on three pillars: protecting young people online, respecting the role of families in digital environments and enabling youth to explore AI technologies responsibly.

According to Google, safety policies prohibit harmful content, including material linked to child exploitation, violent extremism and self-harm, while additional restrictions target age-inappropriate topics.

Safeguards are integrated throughout the AI development lifecycle, from user input to model responses. Systems use specialised classifiers to detect potentially harmful queries and prevent inappropriate outputs.

These protections are also applied to models such as Gemini, which incorporates defences against prompt manipulation and cyber misuse.

Beyond preventing harm, Google aims to support responsible AI adoption through educational initiatives.

Resources designed for families encourage discussions about responsible technology use, while tools such as Guided Learning in Gemini seek to help students explore complex topics through structured explanations and interactive learning support.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!