Network security

Grok chatbot leaks spark major AI privacy concerns

Private conversations with xAI’s chatbot Grok have been exposed online, raising serious concerns over user privacy and AI safety. Forbes found that Grok’s ‘share’ button created public URLs, later indexed by Google and other search engines.

The leaked content is troubling, ranging from questions on hacking crypto wallets to instructions on drug production and even violent plots. Although xAI bans harmful use, some users still received dangerous responses, which are now publicly accessible online.

The exposure occurred because search engines automatically indexed the shareable links, a flaw echoing previous issues with other AI platforms, including OpenAI’s ChatGPT. Designed for convenience, the feature exposed sensitive chats, damaging trust in xAI’s privacy promises.

The incident pressures AI developers to integrate stronger privacy safeguards, such as blocking the indexing of shared content and enforcing privacy-by-design principles. Users may hesitate to use chatbots without fixes, fearing their data could reappear online.

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

Google urges users to update Chrome after V8 flaw patched

Google has patched a high-severity flaw in its Chrome browser with the release of version 139, addressing vulnerability CVE-2025-9132 in the V8 JavaScript engine.

The out-of-bounds write issue was discovered by Big Sleep AI, a tool built by Google DeepMind and Project Zero to automate vulnerability detection in real-world software.

Chrome 139 updates (Windows/macOS: 139.0.7258.138/.139, Linux: 139.0.7258.138) are now rolling out to users. Google has not confirmed whether the flaw is being actively exploited.

Users are strongly advised to install the latest update to ensure protection, as V8 powers both JavaScript and WebAssembly within Chrome.

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

EU and Bangladesh strengthen cooperation on cybersecurity and digital economy

The EU has engaged in talks with the Bangladesh Telecommunication Regulatory Commission to strengthen cooperation on data protection, cybersecurity, and the country’s digital economy.

The meeting was led by EU Ambassador Michael Miller and BTRC Chairman Major General (retd) Md Emdad ul Bari.

The EU emphasised safeguarding fundamental rights while encouraging innovation and investment. With opportunities in broadband expansion, 5G deployment, and last-mile connectivity, the EU reaffirmed its commitment to supporting Bangladesh’s vision for a secure and inclusive digital future.

Both parties agreed to deepen collaboration, with the EU offering technical expertise under its Global Gateway strategy to help Bangladesh build a safer and more connected digital landscape.

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

Hackers steal data from over a million Allianz customers

A data breach at Allianz Life exposed the personal information of around 1.1 million customers, including names, addresses, and dates of birth.

Hackers accessed a customer database hosted on Salesforce, stealing emails, phone numbers, and in some cases, Social Security numbers.

The company confirmed the breach in late July but has not specified the full scale of the incident while its investigation continues.

Cybercrime group ShinyHunters is believed to be behind the attack and is reportedly preparing a data leak site to extort victims.

Several global companies using Salesforce infrastructure, including Qantas and Workday, have reported similar incidents linked to the same hacking collective.

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

Google sees growing demand for cloud data sovereignty

Google’s Cloud Experience lead Hayete Gallot says developer interest in sovereign cloud solutions is rising sharply amid AI concerns. More clients are asking to control how and where their data is stored, processed, and encrypted within public cloud environments.

Microsoft said it could not guarantee full cloud data sovereignty in July, increasing pressure on rivals to offer stronger protections.

Gallot noted that sovereignty is more than location. Cybersecurity measures such as encryption, ownership, and administrative access are now top priorities for businesses.

On AI, Gallot dismissed fears that assistants will replace developers, saying skills like prompt writing still require critical thinking.

She believes modern developers must adapt, comparing today’s AI tools to learning older languages like Pascal or Fortran.

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

Anthropic introduces a safety feature allowing Claude AI to terminate harmful conversations

Anthropic has announced that its Claude Opus 4 and 4.1 models can now end conversations in extreme cases of harmful or abusive user interactions.

The company said the change was introduced after the AI models showed signs of ‘apparent distress’ during pre-deployment testing when repeatedly pushed to continue rejected requests.

According to Anthropic, the feature will be used only in rare situations, such as attempts to solicit information that could enable large-scale violence or requests for sexual content involving minors.

Once activated, Claude AI will be closed, preventing the user from sending new messages in that thread, though they can still access past conversations and begin new ones.

The company emphasised that the models will not use the ability when users are at imminent risk of self-harm or harming others, ensuring support channels remain open in sensitive situations.

Anthropic added that the feature is experimental and may be adjusted based on user feedback.

The move highlights the firm’s growing focus on safeguarding both AI models and human users, balancing safety with accessibility as generative AI continues to expand.

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

WhatsApp trials AI-powered Writing Help for personalised messaging

WhatsApp is testing a new AI feature for iOS users that provides real-time writing assistance.

Known as ‘Writing Help’, the tool suggests alternative phrasings, adjusts tone, and enhances clarity, with all processing handled on-device to safeguard privacy.

The feature allows users to select professional, friendly, or concise tones before the AI generates suitable rewordings while keeping the original meaning. According to reports, the tool is available only to a small group of beta testers through TestFlight, with no confirmed release date.

WhatsApp says it uses Meta’s Private Processing technology to ensure sensitive data never leaves the device, mirroring privacy-first approaches like Apple’s Writing Tools.

Industry watchers suggest the new tool could give WhatsApp an edge over rivals such as Telegram and Signal, which have not yet introduced generative AI writing aids.

Analysts also see potential for integration with other Meta platforms, although challenges remain in ensuring accurate, unbiased results across different languages.

Writing Help could streamline business communication by improving grammar, structure, and tone accuracy if successful. While some users have praised its seamless integration, others warn that heavy reliance on AI could undermine authenticity in digital conversations.

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

Bragg Gaming responds to cyber incident affecting internal systems

Bragg Gaming Group has confirmed a cybersecurity breach affecting its internal systems, discovered in the early hours of 16 August.

The company stated the breach has not impacted operations or customer-facing platforms, nor compromised any personal data so far.

External cybersecurity experts have been engaged to assist with mitigation and investigation, following standard industry protocols.

Bragg has emphasised its commitment to transparency and will provide updates as the investigation progresses via its official website.

The firm continues to operate normally, with all internal and external services reportedly unaffected by the incident at this time.

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

AI toys change the way children learn and play

AI-powered stuffed animals are transforming children’s play by combining cuddly companionship with interactive learning.

Toys such as Curio’s Grem and Mattel’s AI collaborations offer screen-free experiences instead of tablets or smartphones, using chatbots and voice recognition to engage children in conversation and educational activities.

Products like CYJBE’s AI Smart Stuffed Animal integrate tools such as ChatGPT to answer questions, tell stories, and adapt to a child’s mood, all under parental controls for monitoring interactions.

Developers say these toys foster personalised learning and emotional bonds instead of replacing human engagement entirely.

The market has grown rapidly, driven by partnerships between tech and toy companies and early experiments like Grimes’ AI plush Grok.

At the same time, experts warn about privacy risks, the collection of children’s data, and potential reductions in face-to-face interaction.

Regulators are calling for safeguards, and parents are urged to weigh the benefits of interactive AI companions against possible social and ethical concerns.

The sector could reshape childhood play and learning, blending imaginative experiences with algorithmic support instead of solely relying on traditional toys.

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

Fake Telegram Premium site spreads dangerous malware

A fake Telegram Premium website infects users with Lumma Stealer malware through a drive-by download, requiring no user interaction.

The domain, telegrampremium[.]app, hosts a malicious executable named start.exe, which begins stealing sensitive data as soon as it runs.

The malware targets browser-stored credentials, crypto wallets, clipboard data and system files, using advanced evasion techniques to bypass antivirus tools.

Obfuscated with cryptors and hidden behind real services like Telegram, the malware also communicates with temporary domains to avoid takedown.

Analysts warn that it manipulates Windows systems, evades detection, and leaves little trace by disguising its payloads as real image files.

To defend against such threats, organisations are urged to implement better cybersecurity controls, such as behaviour-based detection and enforce stronger download controls.

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