After a deepfake voice fooled her grandfather, this founder sprang into action
After her grandfather was scammed by a deepfake of his brother's voice, Tarini Padmanabhuni founded DetectifAI, a San Francisco startup building AI models small enough to run directly on smartphones and flag fake voices in real time, and it's now one of the companies competing in Startup Battlefield at TechCrunch Disrupt.
The story of DetectifAI begins with a single, chilling phone call. A grandfather hears his brother's voice, pleads for help, and sends money. The voice is a fabrication, a deepfake so precise that it bypasses the most intimate form of trust we have. Tarini Padmanabhuni turned that family betrayal into a startup, and now DetectifAI is stepping onto the TechCrunch Disrupt stage.
The company's pitch is elegantly simple: run the detection model on the device itself. No cloud round-trip, no latency, no privacy leak. The AI listens to the voice in real time and flags the telltale artifacts of synthesis. It is a technical choice that also happens to be a strategic one. By keeping the model small and local, DetectifAI sidesteps the infrastructure costs that sink many AI startups and positions itself as a privacy-first solution in a market where trust is the product.
The timing is not accidental. Voice deepfakes have moved from novelty to weapon. They are used to impersonate executives, defraud the elderly, and manipulate elections. The Federal Communications Commission has already ruled that AI-generated voices in robocalls are illegal, but enforcement lags behind the technology. DetectifAI is not trying to solve the legal problem; it is building the technical tripwire that makes the scam fail at the moment of contact.
For the remote work economy, the implications are quieter but no less real. As distributed teams rely on voice calls and video meetings, the authenticity of a voice becomes a security boundary. A deepfake voice could authorize a wire transfer, approve a contract, or extract sensitive information from a remote employee. DetectifAI's on-device approach could become a standard layer of verification for any organization that operates across distances, where the face on the screen is not always the person you think it is.
The company is still early, and the Startup Battlefield is a proving ground, not a finish line. But the origin story gives it a moral clarity that many AI ventures lack. This is not a solution in search of a problem. It is a response to a wound, engineered into a defense. The market will decide whether DetectifAI scales, but the need it addresses is not hypothetical. It is already happening in living rooms and boardrooms, one convincing voice at a time.