Blog

For years, websites have relied on technology like CAPTCHA to prove that the user on the other end is human. More recently, social media platforms have turned to photos or videos to verify users’ identities. Now, a startup called Helix is betting that voice technology can do the same job, without requiring users to hand over sensitive personal information.

The voice technology, which can run directly on a user’s device, has attracted backing from Siri co-founder Dag Kittlaus, who will also act as an advisor.

The company raised $6 million in funding in a round led by Chicago Ventures, with participation from BAM Ventures, Aglaé Ventures, Corazon Capital, Tusk Ventures, Yonder Ventures, Rule30, and Brown Alumni Group.

The startup was founded by Alex Oberg, Anna Dorofiyenko, and Mike Hall in 2025. Oberg previously worked at Ticketmaster, which acquired his startup; Dorofiyenko led data science teams at Live Nation and Zillow; and Hall has worked at Nike and Tinder, where he helped develop the initial matching algorithm.

At Ticketmaster, Oberg played a role in creating Verified Fan, which is designed to keep bots out of the platform. As AI advanced, he started thinking along the same lines for a new solution.

“As AI started to come up in the last two years, I was realizing that it’s going to make it increasingly difficult to determine who’s human and who’s not on the Internet. Identity on the Internet’s always been about proxies for identity, right? It’s your email, it’s your phone, it’s your IP address, it’s your device fingerprint, it’s your device ID,” Oberg said.

He noted that with the rise of AI agents like Instinct and Muse, the industry needs to rethink human identity on the internet and rely on biometric signals like voice, which is where Helix comes in.

Its core technology can run directly on a device to determine whether someone is human and estimate their age. The company is also running pilots to explore whether voice can be used a biomarker to detect conditions such as adrunkenness or distress.

The company can deploy its technology to replace or supplement CAPTCHA and help block bot attacks. Helix said that because its model can run locally even on an iPhone 13, the cost of deploying the technology is minimal. The startup said one of its customers is Zeel, an in-home healthcare provider that uses Helix’s tech to verify the identities of providers visiting customers’ homes.

Oberg said verifying workers, such as cab drivers, is another use case the startup is exploring. He also said that age-verification regulations for social media platforms around the world are forcing users to trade privacy for access. Helix’s voice technology, he said, could offer an alternative.

“There’s like an avalanche of regulation coming down because essentially society has decided children should not be able to use the Internet as adults. That has created a bit of a strawman argument that everyone needs to trade their privacy for the sake of security and child safety. We are that company that’s saying, nope, don’t agree. It’s a four-second voice challenge,” he said.

The CEO also emphasized the limitations of Helix’s technology, noting that it cannot reliably detect whether someone is lying. He added that its voice technology doesn’t identify individual speakers and claimed that even if the company stores a voiceprint, it cannot be used to reconstruct a person’s voice.

AI-powered fraud is also a growing problem in phone calls. But Helix believes its technology isn’t yet capable of reliably detecting AI agents during live calls.

Siri co-founder Kittlaus said that voice identification technology has finally reached a point where it works reliably, positioning Helix to scale as AI-powered fraud increases. He also said that Helix’s work to compress its models gives the startup a cost advantage.

“What folks at Helix are doing is, instead of teaching computers how to understand speech, these guys are understanding, what the speech of the speaker actually means, what’s going on with them, who they are they lying, are they impaired? That side of it is fascinating to me within the larger picture that I just described,” he told TechCrunch.

The incoming Democratic majority in Congress appears to be eager to reel in an AI industry that’s been allowed free rein under Donald Trump’s second administration.

On top of a slew of regulations meant to address data center electricity use and AI product liability, Democrats are reportedly eying additional preventative regulations aimed at curbing the tech industry’s excesses before they cause harm.

In a post on social media, Arizona Democratic representative Yassamin Ansari urged that “top of the Democratic agenda on day one needs to be a pause on advanced AI development, regulation of this out of control industry, and a ban on artificial superintelligence.”

Democratic house leader Ted Lieu likewise jumped on the wagon, blasting Republicans during a post-election press conference for letting the “AI industry run wild.”

As the rhetoric flies, a bill being pushed by representatives Greg Casar and Alexandria Ocasio-Cortez, along with senator Bernie Sanders, is gaining steam. Called the Ban Superintelligence Act, the legislation would functionally create a permitting process by which a new cabinet called the Department of Artificial Intelligence would dish out credentials allowing AI labs to work on “advanced” AI.

According to the language of the draft bill, the regulation would require AI labs to give the secretary of the AI agency complete access to their digital and physical infrastructure, including data centers. The secretary’s office would have total authority to define “advanced” AI, with a rolling threshold that can be adjusted as the tech advances.

In addition, the bill would introduce massive penalties for “rogue actors” that fail to comply with the department’s charter, including 20 year prison sentences for the worst offenders and 10 year bans on individual employees from working in the industry altogether.

Whether more moderate Democrats will get behind the Ban Superintelligence Act remains to be seen. But if passed, it would no doubt have major implications for an AI industry which has so far gone unchecked by any higher authority in the United States.

More on politics: Flock Safety Found Handing Out Fat Sacks of Cash to Friendly Politicians

TL;DR: HoverAir launched the Versa, a 2-in-1 hybrid camera that converts from a 163g 3-axis handheld gimbal into a 230g autonomous flying camera in seconds. It is available to order now.

Packing for a hike or weekend trip usually means deciding which piece of gear you can fit into your bag. If you want buttery walking footage or close-up selfies, you reach for a handheld gimbal; if you want third-person follow shots and sweeping views from above, you end up hauling a separate drone, a dedicated flight controller, and extra batteries.

The HoverAir Versa combines both setups into a single pocket-sized rig that transforms between handheld and aerial modes in seconds.

In your hand, the device weighs 163 grams and relies on a 3-axis mechanical gimbal to keep first-person trail footage and vlogs smooth without cropping your video frame. Built around a 1/1.3-inch sensor, it captures up to 17.5 stops of dynamic range, 4K video at 60 fps, 4K HDR at 30 fps, and native 4K vertical clips for social feeds.

Advanced low-light image processing and 10-bit H-Log profiles help keep high-contrast scenes — from mid-day mountain glare to neon-lit city streets — from blowing out highlights or losing subtle shadow details.

When you want an overhead view, you snap on the 67-gram flight kit. With two clicks, it becomes a 230-gram autonomous flying camera that launches and lands on your palm. It tracks and films you at speeds up to 36 km/h (about 22 mph), with no piloting experience required, freeing you to focus on the trail or the conversation.

Credit: HoverAir

If you want more creative angles, the device packs over 10 smart shooting modes, including Dolly Zoom, Orbit, Inception Mode, and a 3D Worlds feature that flies a 360-degree loop to map and reconstruct surroundings in three dimensions.

Piloting remains flexible whether you prefer to go hands-free or fly manually. You can tweak compositions mid-air using simple hand gesture controls, use the optional 35-gram MiCo Mini Controller for quick one-handed adjustments, or pair the device with Beacon and JoySticks for traditional stick control.

The HoverAir Versa is available in black and white starting on HoverAir’s web store, Amazon, and select retail partners.

Replying to MM maintainer Andrew Morton, who asked whether a “Nacked-by” was allowed and told him to “hang in there,” Stoakes described how the problem grew: “Of course the AI wave has been horrible, and is getting significantly worse. At first it was once in a while that I’d see slop. Then it was at least once a week. Then every few days. Nowadays it’s pretty much half of what I see, if not more” ^[2]. He called the change “mostly about cutting down the noise in my inbox” and said he will likely still review other work when needed ^[2][3]. MM maintainer David Hildenbrand acked the series with “Yeah, it’s crazy” ^[5], and reviewer Lance Yang wrote that “with the AI flood these days, inbox zero just isn’t happening” ^[3].

Stoakes had raised the issue before. In July he complained that “the amount of time wasted on AI slop is already too much and we’re only at the start of this,” and about a month ago said “what kills the project is having 1,000 slop submissions that 2 or 3 people have to do something with,” Tom’s Hardware reported ^[4].

The step-back lands the same week Linus Torvalds wrote that “much of” Linux 7.4-rc2 “was driven by various AI fixes” and relayed a BPF maintainer’s view that fixes now outpace new bug reports (MLQ.ai coverage). The kernel does not require AI-assisted patches to be labeled as such ^[4], so the filtering falls to the small group of human reviewers in each subsystem.