WhatsAppMeta

Meta unveils Scam Alert for encrypted WhatsApp chats; optional AI tool flags scam messages on-device without giving Meta access to conversations

Meta has introduced a Scam Alert feature for WhatsApp, currently in beta testing, which uses on-device AI to flag potential scam messages without accessing users' private chats. The tool aims to protect over 3 billion users from scams, including romance-baiting tactics that cost Americans $425 million last year.

CNET13 August 2026 · 23:19 UTC
CuriousCats Full Story

Meta has unveiled a Scam Alert tool for WhatsApp, designed to combat scams by utilizing an on-device machine learning model that flags suspicious messages without compromising user privacy.123

The feature, currently in limited beta, operates by analyzing communication patterns associated with scams, such as requests for personal information and urgent tones.

Users can enable the Scam Alert tool through WhatsApp's settings menu, allowing the app to classify incoming messages from unknown senders.

When a potential scam is detected, WhatsApp displays a discreet banner warning the recipient. Users can then choose to block the sender, report the conversation, ignore the warning, or trust the unknown sender.

“Since this is an AI tool we’re talking about, it’s constantly learning,” a spokesperson noted, emphasizing that users can share messages from trusted contacts to enhance the model's accuracy.

With over 3 billion users globally, WhatsApp aims to protect its community from scams, which have led to significant financial losses. In the past year alone, Americans lost $425 million to scams on the platform.1011

While the Scam Alert feature is still in beta, Meta has not yet announced a timeline for its full public launch, but it represents a significant step in enhancing user security on the world's most popular messaging app.

Key Insight
“The feature is in limited beta, and Meta hasn't announced a full public launch timeline. Users can share the last five messages from a trusted sender to help the model learn, and Americans lost $425 million on WhatsApp in a recent year.”
CuriousCats studied:
1
CNET
“WhatsApp is cracking down on scammers with a new Scam Alert tool.”
CNET →
Ask CuriousCats
What is the Scam Alert feature?
Who developed the Scam Alert tool?
Why is privacy a concern with AI?
Are other messaging apps implementing similar features?
How does the AI flag messages compared to human intervention?
CuriousCats
RELATED COVERAGE:
🇮🇳13 August 2026 · 23:16 UTC
BJP questions Meta over Instagram's 'boost' for Jantar Mantar, 'mute' for Jharkhand; Mallika Bajaj flags 'cognitive warfare' in algorithm

Bajaj argues algorithmic amplification is not just software rules, citing virality, engagement, bots, and platform architecture as influencers. She calls for greater awareness among young users about how social media platforms operate, amid NDTV's data investigation into the visibility gap.

Xiaomi unveils HyperOS 4 with glass design aesthetics, improved features; reveals first 26 devices for Android 17 beta
91mobiles.com🇮🇳13 August 2026 · 17:13 UTC
Xiaomi unveils HyperOS 4 with glass design aesthetics, improved features; reveals first 26 devices for Android 17 beta

The beta program rolls out in three phases starting August 14, with the first batch including the Xiaomi 17 series, Redmi K90 models, and Pad 8 tablets. Xiaomi claims system response time improves by up to 17.5% and app stability by 28.9%.

Meta adds AI screening to detect WhatsApp scams; optional feature uses on-device machine learning to flag suspicious messages
🇮🇳13 August 2026 · 18:17 UTC
Meta adds AI screening to detect WhatsApp scams; optional feature uses on-device machine learning to flag suspicious messages

The feature is optional, meaning users can choose to enable it. It relies on on-device machine learning, which processes messages locally on the device rather than sending them to servers, potentially enhancing privacy while flagging suspicious content.

Mark Zuckerberg's 6,500-word AI manifesto criticized as vague and disconnected from reality; former Meta civil rights lead questions which Zuckerberg wrote it
Tech Policy Press🇮🇳13 August 2026 · 21:31 UTC
Mark Zuckerberg's 6,500-word AI manifesto criticized as vague and disconnected from reality; former Meta civil rights lead questions which Zuckerberg wrote it

The author, who led a team building a mandatory training module on civil rights, wonders how many such teams remain after Zuckerberg's 2023 memo mandated leaner non-engineer groups and laid off thousands. He also notes that AI-driven prosperity is concentrated in the top 1%, who hold as much wealth as the bottom 90%.

GitHub's Secure Open Source Fund invested over $500,000 across 50 projects; AI helps maintainers investigate, prioritize, and respond faster, but maintainers still provide context, judgment, and accountability.
The Next Platform🇮🇳13 August 2026 · 22:03 UTC
GitHub's Secure Open Source Fund invested over $500,000 across 50 projects; AI helps maintainers investigate, prioritize, and respond faster, but maintainers still provide context, judgment, and accountability.

Each project received $10,000 USD, with $6,000 during the sprint and $2,000 at six- and 12-month security check-ins. Maintainers turned support into concrete improvements, from hardening CI setups to developing custom Advanced Security Configurations for hundreds of repositories.

Twitch lets streamers opt out of Amazon AI training; setting buried in Security and Privacy, toggling off won't affect captions or safety tools
Mashable🇮🇳13 August 2026 · 12:06 UTC
Twitch lets streamers opt out of Amazon AI training; setting buried in Security and Privacy, toggling off won't affect captions or safety tools

The opt-out option is tucked under Settings > Security and Privacy, described as allowing channel content to train Amazon's generative AI models. Twitch notes that disabling it won't impact AI-supported features like captions and safety tools, but enabling it could improve speech-to-text models across Amazon.

WhatsApp tests on-device Scam Alert that flags suspicious messages privately; users can block, report, or mark chats as trusted
Digital Watch Observatory🇮🇳13 August 2026 · 10:35 UTC
WhatsApp tests on-device Scam Alert that flags suspicious messages privately; users can block, report, or mark chats as trusted

The feature runs entirely on-device, with message content never sent to Meta or third parties, and users can optionally share the last five messages for improvement. WhatsApp also plans to expand its Bug Bounty program to cover the privacy architecture and model behavior.

WhatsApp, Netflix found on phone in rape convict Prajwal Revanna's Bengaluru jail cell; CCB seizes Android 5G, FIR filed
News18🇮🇳13 August 2026 · 08:12 UTC
WhatsApp, Netflix found on phone in rape convict Prajwal Revanna's Bengaluru jail cell; CCB seizes Android 5G, FIR filed

The phone was hidden under Revanna's bed and also contained Instagram and Prime Video, with both Revanna and fellow inmate Pratap Rai denying knowledge of the passwords. Jail authorities suspended Assistant Superintendent Eeranna Rangapur and issued a show-cause notice to the superintendent, while a fresh FIR was registered under the Karnataka Prisons (Amendment) Act, 2022.

Telangana govt warns of fake SIR messages stealing voters' bank details; TGCSB advisory says Election Commission charges no fee for updates
Telangana Today🇮🇳13 August 2026 · 19:45 UTC
Telangana govt warns of fake SIR messages stealing voters' bank details; TGCSB advisory says Election Commission charges no fee for updates

The scam involves fraudsters demanding a nominal ₹5 payment via WhatsApp links to verify voter details, then seeking sensitive banking credentials like CVV and ATM PIN. Authorities urge citizens to report such messages to the nearest police station or cybercrime helpline.

Tata Sons Chairman N Chandrasekaran resigns; exit raises fresh concerns over Trusts' grip on conglomerate, with investors worried about stability
🇮🇳13 August 2026 · 18:30 UTC
Tata Sons Chairman N Chandrasekaran resigns; exit raises fresh concerns over Trusts' grip on conglomerate, with investors worried about stability

The resignation stems from disagreements with Tata Trusts, which controls 66% of Tata Sons, over mounting losses at Air India and other strategies. Tata group stocks lost $4.6 billion in market value on the news, and investors with $54–56 billion exposure are watching succession plans closely.

Get your CIA-level briefing,
in real time.
CuriousCats monitors the internet every minute for you and brings you the most personalized brief of videos, social media posts, news and more.
Download the App
One story brought you here.
CuriousCats brings you everything else worth knowing.
Get CuriousCats