The subtle nuances of a person’s voice – intonation, cadence, and word choice – can reveal far more than previously understood, including details about their education, emotional state, and even financial standing. Now, a new study indicates this information could be exploited through voice-to-text technology, potentially leading to price gouging, unfair profiling, harassment, and stalking.
Published November 19, 2025, in the journal Proceedings of the IEEE, the research highlights growing concerns about the privacy implications of increasingly sophisticated speech-to-text analysis. While humans readily recognize cues like fatigue or nervousness, computers can process vocal data with greater speed and identify a wider range of characteristics, including personal politics and potential health conditions.
“If a corporation understands your economic situation or needs from your voice, for instance, it opens the door to price gouging, like discriminatory insurance premiums,” explains Tom Bäckström, an associate professor of speech and language technology at Aalto University and lead author of the study. The potential for misuse extends beyond financial exploitation. Details revealed through voice analysis – emotional vulnerability, gender, and other personal attributes – could allow cybercriminals and stalkers to identify and track victims across platforms, exposing them to extortion or harassment.
The risks stem from the subconscious nature of these vocal cues. “These are all details we transmit subconsciously when we speak and which we unconsciously respond to before anything else,” Bäckström said in an interview with Live Science. Jennalyn Ponraj, founder of Delaire, a futurist firm focused on the human nervous system and emerging technologies, emphasized the physiological aspect of listening. “Remarkably little attention is paid to the physiology of listening. In a crisis, people don’t primarily process language. They respond to tone, cadence, prosody, and breath, often before cognition has a chance to engage,” she told Live Science.
Bäckström acknowledges the technology is not yet widely deployed, but warns that the necessary tools are already available. “Automatic detection of anger and toxicity in online gaming and call centers is openly talked about. Those are useful and ethically robust objectives,” he said. “But the increasing adaptation of speech interfaces towards customers – so the speaking style of the automated response would be similar to the customer’s style – tells me more ethically suspect or malevolent objectives are achievable.” He admits he is unaware of any documented cases of misuse, but cautions that this could be due to a lack of detection rather than a lack of activity.
The proliferation of voice recordings – voicemails, customer service calls, voice memos – further exacerbates the problem. These recordings create a vast digital record of our voices, comparable in volume to our broader digital footprint. The question remains, Bäckström poses, what will prevent insurers or other entities from leveraging this data to selectively price services based on information gleaned from our voices?
Bäckström expressed concern that even discussing these vulnerabilities could alert potential adversaries. “The reason for me talking about it is because I see that many of the machine learning tools for privacy-infringing analysis are already available, and their nefarious use isn’t far-fetched,” he said. He remains hopeful that awareness and proactive measures can mitigate the risks, warning that without them, “big corporations and surveillance states have already won.”
Researchers are beginning to explore potential safeguards. The Security And Privacy In Speech Communication Interest Group has been established to quantify the information contained within speech, a crucial first step in developing protective measures. One approach involves transmitting only the essential information needed for a transaction, converting speech to text and stripping away identifying vocal characteristics. “The information transmitted to the service would be the smallest amount to fulfill the desired task,” Bäckström explained.
Bäckström believes speech technology holds significant promise, but emphasizes the importance of ethical considerations and robust safeguards. “I’m convinced speech interfaces and speech technology can be used in very positive ways. A large part of our research is about developing speech technology that adapts to users so it’s more natural to use,” he said. “Privacy becomes a concern because such adaptation means we analyze private information—the language skills—about the users, so it isn’t necessarily about removing private information, it’s more about what private information is extracted and what it’s used for.”
Keumars Afifi-Sabet, a technology editor, notes that this research underscores a broader trend of eroding privacy in the digital age. “Studies like this, however, show we’ve barely scratched the surface when it comes to how we can be targeted — especially with something so intimate and personal to us as our own voice,” he wrote. “Although consumer privacy has been massively undermined in the last few decades, there’s plenty room left to use what we hold close to us to be commodified, at best, or in the worst cases, weaponized against us.”
