By Sara Miller, NoCamels -
Miniscule, almost indiscernible behaviors can convey more about our state of mind and inner thoughts than we may realize and are hard to control when interacting with other people.
Using proprietary technology, Petah Tikva-based startup Revealense analyzes these involuntary behaviors, such as facial movements, voice pitch and heart rate, offering insights for organizations who want to know what is going on behind a person’s eyes.
“We analyze human behavior to help other people make the best decisions about humans,” Revealense CPO Amit Cohen tells NoCamels.
Cohen says such technological capabilities are “critical” for two sectors – homeland security and mental health care. Both of which are reliant on careful and precise understanding of a person’s inner thoughts, albeit for very different reasons.
“There’s a lot of similarity between homeland security and mental care,” Cohen explains. “You want to catch the bad guys and you want to help the good guys who need help. And in these two worlds, you have to be very, very accurate – there’s no room for mistakes.”
Founded in 2022, the startup worked on its technology for two years before emerging from stealth very recently with a clutch of clients already in place. And Cohen says that because they are all connected to the security arena, he cannot name them.
He says that the company began with a focus on mental health and homeland security not only because they were the sectors that displayed the most interest in the work they were doing, but also because once the tech was established in these two very demanding fields, it would be simpler to expand to less complex and consequential arenas.
With that in mind, about half of the Revealense team are trained psychologists, he says, alongside a retired Israel Defense Forces general who is the advisor for the security applications of the platform.
Cohen says that unlike other platforms that only analyze involuntary facial movements to understand people’s thoughts and feelings, Revealense takes a more holistic approach that creates a deeper comprehension.
“It’s not only micromovements, it’s all human factors together that we’re analyzing,” he explains.
“We’re trying to create a correlation between all of them [and] we don’t rely only on a single human factor. This is the wrong way to go, because then it can be subject to manipulation or less accuracy.”
Such a narrow approach also fails to take into account ethical considerations such as a subject’s personal history or culture, which Cohen says are very important to Revealense.
He acknowledges that AI and ethics do not always go hand in hand, but explains that the startup brings them together through a policy of what it calls “responsible AI.”
“[This] is one of the main things that is the foundation of our product, because you cannot evaluate human behavior without specific elements that responsible AI brings to the table, for example your background, which affects who you are,” Cohen says.
The platform uses videos – a media and communication tool that has become increasingly popular in recent years – to determine what Cohen calls “the cognitive stress and the emotional stress” of its subjects.
“As we move forward to more digital services, the interaction between people is based on digital interactions, [creating] the requirement to understand what people really think and not only what they say,” Cohen explains.
The users of the platform have a dashboard that uses AI to monitor and analyze a range of criteria.
Blood flow and heart rate, for example, are used as an indicator of emotional stress and measured by changes in skin pigmentation. An increase in the number of times a person blinks and changes in voice pitch are also monitored as signs of similar stress.
And while many studies say that small unconscious facial movements known as micro-expressions can reveal our inner thoughts, Cohen actually warns that these can be easily manipulated in a video.
Instead, he says, Revealense looks for even smaller movements known as facial leaks, which he calls “the most powerful element” of the human face.
These leaks, he explains, are the ones that actually create micro-expressions, and when analyzing a video, it is possible to spot hundreds of facial muscles moving from frame to frame.
Because it focuses on involuntary responses, the platform is adept at spotting so-called deep fakes – advanced video technology that can mimic a person so well it is almost impossible to tell the real from the manufactured.
But according to Cohen, it is the lack of the very human involuntary responses measured by Revealense, which the fakes struggle to replicate, that gives them away as fabricated.
Without divulging details out of security and privacy concerns, Cohen says that the Revealense platform has even been used by certain authorities to determine whether videos of political significance have been manipulated.
All of us, he explains, have the same physical responses to situations that cause us stress, regardless of our own mental state or outlook on the world.
“The bottom line is that the parasympathetic nervous system can never lie, can never be deceptive,” Cohen says of one of the mechanisms in the body that controls heart rate and breathing.
“This is something that is inherent in us all.”
