Your phone says “spam” before you answer. How AI spots suspect calls
The Economic Times · ET Special · October 6, 2026 · economictimes.indiatimes.com Read the articleWidely corroborated
None of the 7 figures cited has a source.
Check started on October 9, 2026
Claims
8 examinedOperators received 721.79 billion incoming calls during April-June 2026, with AI systems flagging 22.99 billion as suspected spam, according to the Ministry of Communications.
TRAI's quarterly report of 4 August 2026 gives 721.79 billion incoming calls and 22.99 billion flagged as suspected spam in Q1 FY2026-27; the figures come from TRAI, which the article attributes to the Ministry of Communications.
These systems also flagged 1.44 billion of 836.31 billion incoming SMS messages.
TRAI's Q1 FY2026-27 figures, as reported by TelecomLead on 4 August 2026, give 836.31 billion incoming SMS and 1.44 billion flagged as suspected spam.
TRAI said in September that major operators had already deployed AI systems and has now made AI-based detection part of its anti-spam rules.
TRAI's press release of 18 September 2026 states that major TSPs have already deployed AI/ML-based systems and that new Regulation 21A of TCCCPR 2018 requires operators to identify likely spam numbers and share the information.
If five or more numbers linked to a sender are flagged within 10 days, operators must investigate and begin graded action, including re-checking identity documents, physical verification, barring outgoing services and disconnection for repeated violations.
TRAI's 18 September 2026 release lists exactly this trigger and these graded actions (KYC re-verification, physical verification, barring of outgoing services, disconnection for repeated violations); Daily Pioneer reports the same.
Businesses must declare their use of A2P calls (autodialled, robocalls, pre-recorded or artificial voices) and the numbers involved; calls made without prior declaration are treated as unwanted commercial communications.
TRAI's 18 September 2026 release defines A2P calls this way, requires pre-declaration to the operator with the numbers used, and says undeclared A2P calls will be treated as UCC.
Airtel says it uses more than 250 derived signals to assess spam behaviour.
tele.net (January 2026) quotes Airtel's chief regulatory officer saying the system evaluates more than 250 parameters to detect spam patterns; the individual signals listed in the article were not found in the sources consulted.
In 2024 Truecaller launched AI Call Scanner, which analyses a caller's voice during a call to indicate whether it is human or AI-generated.
MediaNama reports that Truecaller launched its AI voice scanner on 29 May 2024 (article of 30 May 2024) to tell users during a call whether the voice is AI-generated; India availability only for beta users was not found in the sources consulted.
Google's Scam Detection on Pixel devices in India uses on-device Gemini Nano, stores or sends no audio or transcript, and Google says it "is not 100% accurate and scammers constantly change their tactics".
Google's support page states Scam Detection runs on Gemini Nano on Pixel 9 and later, is available in India, stores or sends no conversation audio or transcription, and contains the quoted accuracy caveat.
Signals found: figures attributed to the ministry of communications, while they were published by the regulator trai · service descriptions rely on the companies' own statements
Measurement
92.1 out of 100 → rounded to 9The text measurements are computed by code, not by the model: repeatable counts. How the score is calculated (in Italian)
Help improve this check
An outcome is not a verdict: it grows with new sources, gathers the judgement of the people who read it, and is run again when the page changes. You can request a new assessment with the Pintru service. Whenever possible, Pintru makes the assessment page public, available to everyone.
Analysis requested by a user on 09-10-2026 at 01:35 using Artificial Intelligence
Not an official determination and not binding on the page publisher.