A discussion about the data that drives the recognition service, plus a look at privacy concerns
Both AI and facial recognition are hot topics in the world of tech, with much of that coming from the innovations of companies like Apple and Google that are bringing the technology to scale for a mass audience.
With such big names involved in the industry, it is easy to forget that there are other players driving innovation in these fields, as well.
One such company is Kairos, an AI engine provider for facial recognition services. I had the chance to speak with Brian Brackeen, founder of the company to discuss more about what they do and some of the social and technical implications that come with something as large as AI and facial recognition in regards to privacy concerns.
Check out the interview below.
Care to introduce yourself and your role with Kairos?
I'm Brian Brackeen, Founder and CEO at Kairos -- We're a face recognition AI company, based in Miami.
In just a couple sentences, what problems are you trying to solve?
When you think about it, there's so much unnecessary friction between how people and machines interact; it's become a blight in our daily lives. Experiences are often based on the machines' terms, not ours; they ask a lot of us!. And now businesses are getting smart to this. Consumers are demanding better, frictionless experiences.
So, face recognition can really help by simplifying those interactions without sacrificing accessibility, safety, or privacy.
Facial recognition is seen as 'scary' by a lot of people, what would you say to those people?
I don't disagree. As a technology, it's gotten a bad rep over the years. And, honestly, I think it's the industry's own doing. Early leaders didn't stop to think about the bigger consequences of enabling forms of surveillance or designing algorithms that are -- for all intents and purposes-- racist.
At Kairos, we are committed to fixing ALL of these problems. By focusing on the many positive outcomes the tech can enable, and listening and learning from our customers, consumers and communities we'll ultimately overcome these early challenges.
You'll see us advocating for more regulation; not less. You'll see us holding the industry accountable; not hiding behind profits. I believe we are unique in our thinking here.
Have you received any resistance to your service by those concerned with privacy?
We work with our customers on best practices around consent, data security, and privacy -- which we welcome. I'm often invited on panels, to debate privacy concerns, and we actively seek out 'resistance' so we can better understand how to address concerns. We won't learn anything by being defensive or hiding away from the big talking points. I'm all about extreme transparency, it's the only way.
It's not uncommon for people's first reaction to be "Oh, that's creepy. Nope."-- yet when we take the time to listen and offer our thinking. Generally, those initial concerns are put to rest. That said, some folks are resolute we are not George Orwell's 1984 vision come true.
We're also resolute in our decision not to work with Government agencies or companies that are looking to exploit people's private data. Many are surprised when we reveal we are in favor of anonymity, and our technology is designed to those values.
I will add, we are grateful for the leadership of George Town Center on Privacy & Technology in our space. It's been an honor to collaborate and learn from their research.
What hardships have you faced with building this service?
I'm a first time, minority founder and we're based in Miami. So, I've seen way more nos than yeses from the big West and East Coast VC firms. Mix in an emerging technology, and we haven't made it easy for ourselves, lol. Is that a hardship though? I see it more as an opportunity to prove everyone wrong!
And, candidly, in the early days-- when we were figuring out how to go-to-market-- before our product was really ready for prime time, we had to let people go and even forgo payrolls so we can keep the lights on and the servers running. Hard, hard decisions, yet ones that on reflection have made us who we are today.
So much of AI requires huge data pools to draw info from, can you go into it a bit on how you obtain that data?
We use a combination of public datasets and our own proprietary data to design our initial algorithms. And we don't pool data, in the same way as perhaps a Facebook might. We're more thoughtful, as it's really a means to an end. Sometimes we collaborate with our customers on using their data to enhance our system. Again, we're thoughtful here about who we work with and why.
Beyond the initial design and through the process of actually performing facial recognition, we are continuously learning from each face we analyze to further increase accuracy. And this is done with our anonymous values in mind; so we don't actually know who the face belongs to, rather we abstract the geometry of the facial features, discard the original image, and turn the geometry into a unique string of letters and numbers-- called a 'template'. This template cannot be reverse engineered.
It's a goal of mine to collaborate with communities, all around the world to create a diversity dataset that will ultimately defeat the biases that exist today in many face recognition algorithms.
What kind of businesses can benefit from Kairos?
The reality is practically any business can benefit from it and what we help our customers do is figure out if they should benefit from it/are those benefits impactful. It goes back to our thoughts on what is best for the consumer.
Do we want to help create a safer experience for a user of a banking app? 100%. Identity fraud is a persistent problem. That's where we see the most value being transferred; mapped to the big problems businesses face today. This concept of 1-2-1 verification is appealing to us, those moments between an individual and a [business] interface. How can we make that experience the best it can be?
Anything exciting coming to the service?
We're excited to launch a new blockchain focused product. Kairos has built a first of its kind identity standard for using biometrics on the blockchain. And plans to open it up to the whole world. Our design preserves anonymity while suffocating malicious intent, such as fraud.
Anything you'd like to close with?
To anyone reading, I welcome your thoughts on any of the topics above. We're obsessed with feedback, so don't be shy!