Loading story…
WhatsApp has begun a limited beta rollout of Scam Alert, an optional feature that uses on-device machine learning to flag likely scam messages from non-contacts while preserving end-to-end encryption. Users can decide whether to block, report, continue, or trust a flagged conversation, and can review Scam Alert activity in the app. WhatsApp said message content does not leave the device for classification, and the company has expanded its bug bounty program to cover the new feature, including its model weights and analytics pipeline.
To reduce privacy risks and targeted manipulation, WhatsApp said it verifies model releases through a transparency process that publishes each version’s SHA-256 hash to a third-party append-only ledger before deployment, while model downloads are routed through OHTTP relays with anonymous credentials. The company also said it measures effectiveness through a confidential federated analytics pipeline using Trusted Execution Environments, anonymous aggregation, and differential privacy, and may let users optionally share the last five messages from trusted chats to improve model accuracy. The rollout comes as Signal has also introduced automatic key verification based on key transparency logs and independent auditing to detect unauthorized public key changes in encrypted conversations.
Track how attackers are adapting to this technology.
3 events from the most recent confirmed update back to the earliest known activity.
Signal introduced automatic key verification, a key transparency-based feature designed to detect unauthorized public key changes in encrypted conversations. The system complements manual safety number checks and relies on cryptographically verifiable logs and independent auditors.
WhatsApp said it is expanding its bug bounty program to cover Scam Alert, including the feature's model weights and analytics pipeline. The expansion is intended to allow external researchers to assess the system's security and verify its scam-detection design.
WhatsApp started a limited beta rollout of Scam Alert, an optional feature that uses on-device machine learning to flag suspicious messages from non-contacts without breaking end-to-end encryption. The feature lets recipients block, report, ignore, or trust flagged chats, and keeps classification on the device.
Vulnerabilities, threat actors, malware, products, organizations, breaches, and observables Mallory has linked to this story. Indicator values are masked here and available in full in the app.
Indicator values are masked on this page. See the values in Mallory Domains, IPs, hashes, and URLs are exportable to your SIEM.
Follow how adversaries are adapting to this technology, and where it touches your stack today.
10 references tracked. Mallory keeps watching after this page renders.
infoq.com
Open sourcetechrepublic.com
Open sourceonlinethreatalerts.com
Open sourceghacks.net
Open sourcemkd-cirt.mk
Open sourcecybersecuritynews.com
Open sourcesecurityweek.com
Open sourceengineering.fb.com
Open sourceMap indicators from this story to your assets and identify affected systems in minutes.
Every observed campaign, victim, and pivot linked to actors named in this story.
Malware, exploits, and IOCs connected to the activity described here.
YARA, Sigma, and Snort rules deployed to your SIEM as soon as they’re published.
Get matching new stories delivered to your team as they break — not the next morning.
Ask questions about this story and take action on the answers.
Other stories Mallory is tracking that overlap on entities, sources, or topic labels.
A recent independent study by Incogni evaluated the privacy practices of major messaging apps, ranking platforms such as Discord, Telegram, Snapchat, Facebook Messenger, and WhatsApp based on 18 privacy criteria. The study highlighted that while Discord scored highest for privacy overall, the rankings can shift depending on which criteria are prioritized, and users are cautioned not to make decisions based solely on aggregate scores. The analysis emphasized the importance of understanding how messaging apps handle user data, respond to law enforcement requests, and protect against social engineering attacks. In parallel, Meta has introduced new anti-scam features for WhatsApp and Messenger to bolster user security. These updates include advanced scam-detection tools in Messenger that warn users about suspicious chats and allow for AI-based scam reviews, as well as new warnings in WhatsApp to prevent users from sharing their screens with untrusted contacts. Additional features notify users when they are contacted by unknown individuals or added to group chats by strangers, providing context and safety tips to help users avoid scams. These developments reflect a broader industry trend toward improving both privacy and security in messaging platforms.
Jul 12, 2026
WhatsApp is preparing to introduce unique usernames that would let people message each other without sharing phone numbers, a change positioned as a privacy improvement for group chats, events, and first-time contacts. The feature is designed without a public directory, requiring users to know an exact username, but officials and cybersecurity authorities have warned it could also make impersonation and phishing easier by allowing attackers to register lookalike names resembling trusted individuals, brands, banks, or government bodies. India’s Ministry of Electronics and Information Technology (**MeitY**) has asked WhatsApp to suspend the rollout until it receives government approval, arguing the feature could fuel phishing, so-called digital arrest scams, and fraud targeting public authorities and financial institutions in WhatsApp’s largest market of more than 850 million users. WhatsApp said it plans safeguards including contextual warnings, reserved usernames for legitimate organizations, lookalike protections, contact-initiation limits, anti-enumeration controls, and abuse-detection systems, while security guidance has urged users to verify suspicious contacts through official channels and advised organizations to reserve official usernames early and clearly communicate their legitimate accounts.
Jul 8, 2026
WhatsApp announced a new optional security mode, **“Strict Account Settings,”** designed to reduce exposure to *highly sophisticated attacks*—particularly **mercenary spyware**—by limiting risky functionality when enabled. The feature is expected to roll out in the coming weeks and includes restrictions such as blocking attachments and media from people not in a user’s contact list; it can be enabled via `Settings > Privacy > Advanced`. WhatsApp positioned the change as an additional layer beyond default **end-to-end encryption**, aimed at higher-risk users such as journalists and public-facing figures, and noted its ongoing legal fight with **NSO Group** over the 2019 Pegasus campaign that targeted roughly 1,400 WhatsApp users. The approach mirrors Apple’s **Lockdown Mode**, which similarly reduces attack surface for a small subset of users who may be personally targeted by advanced threats by disabling or constraining features across core services. Apple documents that Lockdown Mode blocks most message attachment types, limits complex web technologies, restricts incoming FaceTime calls to recent contacts, blocks certain Apple service invitations, and removes some photo-sharing metadata—trading usability for stronger protection against targeted exploitation. Together, the updates reflect a broader industry pattern of offering *opt-in, high-friction hardening modes* to mitigate spyware and other highly targeted intrusion techniques.
Jun 29, 2026