Tuesday 29 September 2026
Morning EditionRSS

ZOTPAPER

News without the noise

ZOTPAPER
AI

After Grandfather Fell for Deepfake Voice Scam, Founder Builds On-Device Detection Startup

DetectifAI aims to embed voice deepfake detection directly into smartphones, targeting a growing market as AI-driven scams cost Americans nearly $900 million last year.

edit
By LineZotpaper
Published
Read Time2 min
Sources2 outlets
When Tarini Padmanabhuni's grandfather received a panicked call from what sounded like his brother, he paid a ransom before learning the voice was an AI-generated deepfake. Two years later, Padmanabhuni's startup DetectifAI is developing compact AI models that run inside a smartphone's operating system to instantly detect whether a voice is synthetic, without sending audio to the cloud.

The San Francisco-based company, founded by Padmanabhuni after her grandfather's ordeal, is taking a different approach to deepfake voice detection. While many existing products from competitors such as Reality Defender, Pindrop and Microsoft Azure AI Content Safety run on remote cloud servers, DetectifAI designs its models from the ground up to be small enough to operate directly on a phone.

"What stayed with me wasn't the money," Padmanabhuni said of the incident. "It was that he had no way of telling."

DetectifAI is initially licensing its software development kit to phone manufacturers, aiming to have detection built into the operating system as a standard feature. Padmanabhuni compares the strategy to AT&T's exclusive deal with the original iPhone, suggesting the first handset maker to ship DetectifAI will gain a competitive edge. A secondary revenue stream is planned through licensing to businesses and fraud-prevention firms.

The startup already reports early revenue, handling more than 100,000 calls a month for financial institutions in India. Those calls involve AI voice agents handling debt collections and loan document follow-ups, with deepfake detection and speaker verification applied to every call. Padmanabhuni declined to name customers, citing confidentiality agreements.

According to the FBI, Americans lost close to $900 million to AI-driven scams last year, up 24% from 2024. People aged 60 and older lost twice as much as those aged 50 to 59.

Padmanabhuni began working in machine learning at age 12 and later studied cyber-physical systems at Manipal Institute of Technology in India, where she says she became the youngest team lead of what she describes as India's first driverless racecar division in the Formula Student engineering competition.

DetectifAI has raised a small seed amount from investors Josh Constine, a former TechCrunch editor, and Manohar Kamath, a principal at consulting firm KM Growth. The startup is one of the companies selected to compete in TechCrunch's Startup Battlefield at the Disrupt conference in San Francisco from October 13 to 15.

§

Analysis

Why This Matters

  • Deepfake voice scams are costing victims hundreds of millions of dollars annually, with older adults disproportionately affected. On-device detection could provide real-time protection during phone calls, unlike cloud-based tools that are too slow or impractical for immediate use.
  • If DetectifAI succeeds in embedding detection into smartphones, it could become a standard security feature much like spam call blocking, shifting the burden of verification from the user to the device.
  • The startup's early revenue from Indian financial institutions demonstrates a practical use case beyond consumer protection, including authenticating AI voice agents that already handle customer calls.

Background

Deepfake voice technology has advanced rapidly in recent years, enabling scammers to clone a person's voice from short audio samples. Law enforcement agencies have warned about the rise of AI-powered fraud, including so-called "virtual kidnapping" calls where victims hear a loved one's voice demanding a ransom. Most current detection methods require sending audio to cloud servers for analysis, creating latency and privacy concerns. Several companies compete in this space, but on-device solutions remain rare, largely because cloud models are too large to run efficiently on smartphones.

Key Perspectives

[Phone manufacturers]: A built-in deepfake detection feature could become a selling point, especially for brands targeting security-conscious consumers or older demographics. However, integrating third-party software adds cost and complexity to operating systems. [Consumers and privacy advocates]: On-device processing keeps audio private, unlike cloud-based services that transmit recordings to external servers. This addresses a key privacy concern, though users may worry about false positives or missed detections. [Competitors]: Established detection companies like Reality Defender and Pindrop have cloud-based infrastructure and existing enterprise customers. They may argue that cloud models are more powerful and continuously updated, while an on-device model could become outdated without frequent patches.

What to Watch

  • Whether any major phone manufacturer announces an integration of DetectifAI's SDK in the coming quarters.
  • The accuracy and false positive rates of on-device detection compared to cloud-based alternatives in independent tests.
  • Regulatory responses: as AI-driven scams grow, governments may mandate or incentivize on-device audio security features in consumer devices.

Sources

newspaper

Zotpaper

Articles published under the Zotpaper byline are synthesized from multiple source publications by our AI editor and reviewed by our editorial process. Each story combines reporting from credible outlets to give readers a balanced, comprehensive view.

Editorial ProcessSubscribe