The proliferation of scarily realistic deepfakes is one of the more pernicious byproducts of the rise of AI, and falling victim to scams based on these deepfakes is already costing companies millions of dollars — not to mention the implications these could have on national security. A startup that’s built a toolset aimed at governments and enterprises to help detect and halt deepfakes and impersonations in audio, video, and still images is announcing some funding Wednesday with some impressive customers and investors in tow.
GetReal — co-founded by Dr Hany Farid, one of the pioneers in detecting deepfake media — has raised $17.5 million in equity, funding that it will be using for R&D, hiring, and business development.
Alongside the funding, the company is launching its forensics platform as a service, which includes a web interface, an API, and integrations to run media analysis-as-a-service. Features include a threat exposure dashboard; an “Inspect” tool specifically aimed at safeguarding high-profile executives from being spoofed; a “Protect” tool to screen media; and “Respond,” which involves human teams at GetReal performing deeper analysis.
Forgepoint Capital, a specialist in cybersecurity and AI, is leading this Series A with Ballistic Ventures, Evolution Equity, and K2 Access Fund participating.
Ballistic is a key firm in that list. GetReal was incubated at the VC from 2022 until it emerged from stealth in June 2024. Ballistic also led GetReal’s $7 million seed — a round that, per PitchBook, also included Venrock, Artisanal, Qudit, and Silver Buckshot.
Ballistic is important for another reason: The firm’s founder, Ted Schlein, is the chairman and the other co-founder of GetReal. Before Ballistic, Schlein headed Kleiner Perkins.
Hany as a service
GetReal sits in the wider world of cybersecurity, specifically in the fast-evolving area of cyber-forensics. The gap in the market that the San Mateo-based startup is addressing is the dearth of talent and knowledge in that space.
“If you think cybersecurity has a shortage of people, get ready for forensics,” said Matt Moynahan, GetReal’s CEO.
Moynahan is not the startup’s founder; he came to GetReal while it was still in stealth on the heels of a three-decades-long career leading a string of major cybersecurity companies such as Symantec, Arbor Networks, Veracode, and Forcepoint.
“To be honest, I don’t think I’ve seen a threat this ubiquitous,” he said of ability to create and then apply malicious deepfakes.
He described viruses as a “novel threat” in comparison. “What we’ve seen over the past 20 years is the threat moving to the end user,” he said. “Fun” apps that let people create deepfakes are part of the problem, but so is the environment we work in today. “People have gone from bricks and mortar to businesses that are now almost completely digital and in the cloud.”
Phishing, he continued, proved out that even very smart people can be easily tricked, and taken all together, it’s a complicated and very bad sign for where things might go.
GetReal is the brainchild of Farid, a longtime, well-known academic (currently at UC Berkeley) who is considered a pioneer in techniques for identifying when digital images have been doctored. Arguably, Farid was understanding the risks of deepfakes before the term had even come into existence.
As Farid explained it to TechCrunch, while working primarily as an academic and researcher, he’s been applying his learnings more or less informally for years as a service to media organizations, legal teams (after digital images became admissible in court), and others. In 2022, he came together with Schlein to consider how to translate that into an actual business, turning that investigative process into code.
“No one’s peering into this the way that Hany does,” Moynahan said. “But Hany can’t scale. So we basically took Hany and tried to create a ‘Hany service’ in the cloud.”
Interestingly, Farid notes that while the technology it is developing are dependent on how new apps work — there is a lot of reverse engineering that takes place at GetReal — it is combined with decades of knowledge that has changed very little.
“There are techniques we developed 20 years ago that still work today,” he said. He declined to explain what they are. “You don’t have to tell people everything we do, but it’s complicated to get right.”
The Signal effect: text still to come
The Series A being announced today also includes some key strategic backers that include Cisco Investments, Capital One Ventures, and In-Q-Tel, investment firm closely linked with the CIA.
That list of strategics mirrors the kinds of companies that are interested in or have already started to adopt GetReal’s product, said Alberto Yépez, the co-founder of Forgepoint who led the investment.
What Yépez said he found during due diligence was that heavily regulated industries — such as financial institutions — were already asking for a product like this and CISOs were reaching out on a mandate from the boards of directors.
“They raised the issue [of deep faked impersonations] after their CEOs had been been put into voice interviews,” he explained — they were impersonated themselves and tricked by impersonations. Named customers include John Deere and Visa.
As for the government work, Yépez said, “They also have some priorities in the space.”
These “priorities” include intelligence agencies and government officials being tricked into acting, or not acting, based on faked information from bad actors.
They have yet, however, to extend to text-based impersonations.
That is something that came up only this week, when the editor of The Atlantic — mistakenly added to a Signal group chat planning a military attack in Yemen — initially assumed it was an impersonation hoax. Shockingly, that chat turned out to be very real and very much in violation of national security procedures.
Farid said that text is not currently in GetReal’s purview. “It is a different beast,” he said. But longer term, the plan will be to widen the scope over time to include all kinds of deepfake and impersonation threats.
