A new startup called Deveillance (pronounced dee-veil-ance) announced its first-ever gadget earlier this week: a sleek, portable tabletop orb that aims to jam nearby devices to prevent them from recording voices.
Called Specter I, the microphone jammer is a combination of ultrasonic frequency transmitters and artificial intelligence designed not only to block devices trying to capture someone’s speech, but also to detect and record nearby microphones, all while being small enough to carry around. It’s still in development, but the company plans to sell the Specter I in the second half of 2026 for $1,199.
THE announcement caused a lot of noise on social networks. It has been promoted by some as a cyberpunk-style resistance technology against the ever-growing category of AI wearables that are always listening but also became the target of a storm of skepticism from blue-checked critics of X who were eager to dismiss it as too good to be true.
The specter of revelation I.
Courtesy of Main Irtiza Aftab/Deveillance
“I didn’t expect it to go so viral,” says Aida Baradari, a recent Harvard grad who founded Deveillance and developed Specter I. “I’m grateful to have had the opportunity to work on this. I’m also very grateful, honestly, that people care.”
Baradari was motivated to build this device to counter those constantly listening devices that the AI boom has ushered in, such as the wristband. Amazon property Bee AI or the Friend pendant.
“People should have a choice in what they want to share, especially in conversations,” says Baradari. “If we can no longer converse without fear of saying something that is potentially taken out of context or wrong, then how are we going to make human connections in this new era?
Private time
It is easy to understand why this anxiety about privacy has increased, because government surveillance is popular in the United States. ICE builds its own surveillance systems around everything, from social networks has everyone’s phones to himself employee list. This tension also runs deep within the private sector, as big tech companies ICE fuels while collecting, purchase, and using every bit of your personal data.
In February, when home security camera company Ring launched a Super Bowl commercial on using its cameras to find lost dogs, viewers were appalled by the privacy implications of a neighborhood panopticon and responded immediately repression. This led Ring to backpedaling. A week later, the company announced that it would no longer pursue its efforts planned partnership with the equally controversial security company Herd security.
“People are starting to realize that they may not have a private life at some point,” says the musician and YouTuber. Ben Jordanwhich makes videos about security and privacy issues like audio jammers and Flock security cameras.
Like the amateur developer who created an app to notify people if someone is wear smart glasses nearby, Specter I is another effort to give users a way to regain control of their privacy. But for a device that uses AI and speakers to block other AI and microphones, we must first prove that the technology works. Skeptics say Deveillance’s claims seem far-fetched.
“These are very big promises,” Jordan says. “Unfortunately, they kind of clash with physics.”
Blur
Ultrasonic microphone jammers existed long before the Cold War, built and developed over decades by intelligence agencies and tinkerers. They also became a kind of pocket industrywhere you can buy jammers on sites like Alibaba or create your own using GitHub’s software.
Audio jammers tend to be bulky, thick bricks because the frequency transmitters and power sources take up space. Make the device powerful enough to work, and it’s probably too big to be discreet. Too small, and the jammer will not have the ability to properly disrupt a microphone.
“We are aiming for a light and small device, although this might prove difficult to achieve due to physical constraints,” Baradari wrote in a text to WIRED.
To bridge this gap, Baradari states that the Specter will use AI to alter speech, not just obscure it with a wall of sound. The device sends AI-generated cancellation signals intended to fool automatic speech recognition (ASR) technology. While the plan is to bring the transmitters to a seemingly silent level, the current working version of the Specter I produces an audible hum.
AI is used to target a range of ultrasound frequencies specifically suited to average human voices. “The result is an overall muddled recording, so environmental noises are also muddled,” Baradari wrote. “In traditional jammers, voices can be reconstructed, or the jamming can sometimes be bypassed by ASR systems. With our method, we ensure that this does not happen.”
But Melissa Baese-Berk, a linguistics professor at the University of Chicago, says: “There is so much variation in people’s voices. There is no specific signal that sounds like ‘voice signal’.”
Baradari claims that Specter does not track any voices or what people say, citing that the algorithm is optimized based on Deveillance’s internal development and training. After all, there’s no microphone inside the Specter I. “The AI is optimized to send signals that can’t be reconstructed in post-processing,” she says.
Deveillance also claims that the Specter can find nearby microphones by detecting radio frequencies (RF), but critics say finding a microphone via RF broadcasts is not effective unless the sensor is immediately nearby.
“If you could detect and recognize components via RF as Specter claims, it would literally transform technology,” Jordan wrote in a text to WIRED after build a device to test the detection of RF signatures in microphones. “You would be able to do radio astronomy in Manhattan.”
Deveillance is also investigating ways to integrate nonlinear junction detection (NLJD), a very high frequency radio signal used by security professionals to detect hidden microphones and bugs. NLJD detectors are expensive and used mainly in professional contexts such as military operations.
Even if a device could detect the exact location of a microphone, objects around a room can change the way frequencies propagate and interact. The frequencies transmitted could also be a problem. There have not been adequate studies to show what effects ultrasonic frequencies have on the human ear, but some people and many animals can hear them and find them obnoxious, even painful. Baradari acknowledges that his team needs to do more testing to see how pets are affected.
“They just can’t do that,” said engineer and YouTuber Dave Jones (who runs the EEVblog channel) wrote in an email to WIRED. “They’re using the classic trick of using wording to imply that it will detect all types of microphones, when they’re probably just scanning for Bluetooth audio devices. It’s totally rubbish.” Baradari reiterates that the Specter uses a combination of RF and low energy Bluetooth to detect microphones.
WIRED asked Baradari to share any evidence of Specter’s effectiveness in identifying and blocking microphones near a person. Baradari shared a few short video clips of people putting their phones up to their ears and listening to audio clips – which were likely jammed by the Specter – but these videos don’t actually prove that the device works.
Future imperfect
Baradari accepted the criticism with enthusiasm, acknowledging that the technology is still developing. “I really appreciate these comments because they make me think and see more things as well,” Baradari says. “I believe that with the ideas we have and are integrating into one device, these concerns can be addressed.”
People were quick to mock the Specter I online, calling the technology a cone of silence of Dune. Today, the Deveillance website states: “Our goal is to make the cone of silence a reality. »
John Scott-Railton, cybersecurity researcher at Citizen laboratorywho is critical of the Spectrum I, hailed the virality of the device as an indication of the real thirst for this type of gadgets to reclaim our privacy.
“The positive side of this explosion is that it is a Ring-like moment that highlights how quickly and intensely consumer attitudes have changed toward ubiquitous recording devices,” says Scott-Railton. “We need to create products that do all the cool things people want, but without causing massive violations of privacy and consent. You need device-level controls, and you need regulations of companies that do that.”
Cooper Quintin, a senior technologist at the Electronic Frontier Foundation, echoed these sentiments, although critics say Deveillance’s efforts are flawed.
“If this technology works, it could be a boon for many,” Quintin wrote in an email to WIRED. “It’s nice to see a company creating something to protect privacy instead of working on new and creative ways to mine our data.”
