When Attackers Have AI, Verify the Person Each Time
Episode Summary
Everything good about AI is what makes it hard in security. Kris Bondi, CEO and co-founder of Mimoto, starts with the asymmetry: attackers innovate and automate at scale while enterprises implement slowly, so a step behind can quickly become ten. Her answer is a different question rather than a higher wall. What she calls anomaly detection on steroids tells you an account is behaving oddly. It cannot tell you who is using it. Her company matches patterns at the person level instead, on habits as individual as how somebody builds a Unix command line, and keeps checking after login, which she calls continuous validation. That matters most after a deepfake succeeds: an intruder in an admin account touches what an admin touches, so the anomaly may not show up. She closes on the question to put to any security tool: how adaptive and dynamic is it, and can it act in real time rather than send one more alert.
Key takeaways
Her opening line is one half of the asymmetry in one sentence. Everything good about AI is a part that makes it difficult in security, because everything good about it also lets bad actors change their approach, innovate on it, evolve it, and automate at scale
The second half of that asymmetry is speed, and it is the one enterprises control. Enterprises tend not to implement or innovate as quickly as the people attacking them, so an organization that starts a step behind can quickly end up ten steps behind
The first benefit she names for defenders is a capability many companies do not have at all. A red team tries to breach you and a blue team tries to stop it, some companies run a purple team that does both, and many organizations do not have the ability to have either
The second benefit is inside the process rather than at the perimeter. Running the same analysis internally shows a team how to take touchpoints out of a process and where the checks belong, and she says the automation of reading signals and responding is what was not possible before
She dates her own product to the same shift, and says plainly what it does. Data streams come in, the algorithms are picked according to which streams they are, and a composable model is created on the fly to make a determination and take action in real time, none of which was possible three years earlier
The distinction she stops the conversation to make is the center of the episode. Analyzing what is happening inside systems is anomaly detection on steroids; the pattern matching layered on top of it identifies a person, which is a different job
Stated as the difference between two alerts, it is unmistakable. The anomaly says there may be something odd with Christina’s account, and the person-level match says with certainty that the person using it is somebody else
The capability she is waiting for is intent, and she puts a horizon on it. She expects LLMs to develop to the point where they can be queried about intent, so a system could ask what it looks like someone is trying to do rather than only that the behavior is unusual, and at recording she expected that potentially within the year
On deepfakes she refuses to describe a solved problem. Deepfakes are an arms race, attackers will continue to get better, and detection will have to keep innovating in response
Detection at the door is a category with several members. Outside organizations focus on identifying deepfakes at the authentication step, companies using government-issued IDs often run something else behind the scenes as well, and visual and audio systems look for glitches that suggest the face or the voice is not the right person
Past the door, her word for the second layer is the one that names the control. Continuous validation expands beyond the face to how somebody interacts, which is the other place to catch an intruder
The consequences she lists are not a single risk. On the anti-fraud side deepfakes can cost companies a lot of money; on the other side, a deepfake used to get into a system is where a ransomware attack may happen or data may be stolen, and those often run through an account takeover
Her worked example is why the account matters more than the entry. A systems admin account touches admin things, so an intruder who used a deepfake to get in touches exactly what the real admin would touch, at hours the real admin might also be working
Her one-sentence account of where the standard control fails is the sharpest line in the episode. The anomalies may not show up, because the intruder is touching the things they would normally touch
What replaces the anomaly is a set of habits nobody chooses on purpose. At the person level the signals include how somebody uses a mouse and a keyboard and what they do at a web level, and she declines to list the rest
Her illustration is two people who agree on everything except muscle memory. Her two technical co-founders, one with a computer science degree and one self-taught early, build their Unix command lines differently, which is the kind of thing that is inherently specific to a person
She upgrades the fingerprint metaphor Christina offers rather than accepting it. It is a multidimensional fingerprint, and her illustration is her own left pinky on the caps key, which she immediately disqualifies: she has said it in public often enough that it can never be an identifier of her
The reason she started the company is a loss that has not happened yet. She says society has been lucky that nobody getting into a system unseen has yet caused a larger loss of life or a disruption in how people live, and that this is why the company exists
The idea came from a product that pointed the other way. She had been with an Israeli company that built a neural network predicting what people were doing offline in real time, which showed her what a digital double could do, and she says privacy protection is what her current company fixed this time around
On the Social Security change Christina raises, she moves the problem off the technology. It goes beyond voice, because somebody calling a clerk who knows neither party would not need to clone a voice at all, and her question is whether you have to talk to a person in the first place
The number she puts on moving everybody online instead is the one to remember. A bank will get 200 million bots in a relatively small amount of time trying to set up accounts or make distributions, which she calls a large number but not an unheard of one
Her first piece of advice puts security inside the AI program rather than beside it. Companies on a journey of implementing AI should extend that innovation into their security, rather than trying to use old security for new problems
Her second is a question to ask of any tool, and it is the one to take into a vendor meeting. How adaptive and dynamic is it, meaning can it respond to the problem as presented today and as presented tomorrow
Her third turns the answer into a test the buyer can apply. Real time has to mean taking action, not sending an alert, because a tool that only alerts is putting more of a tidal wave on top of the team it was bought to help
About Kris Bondi
Kris Bondi is CEO and co-founder of Mimoto, a security company that identifies people rather than credentials, matching patterns at the person level so a system can tell that the person using an account is not its owner. She works with two technical co-founders. Before Mimoto she was with an Israeli company that built a neural network to predict what people were doing offline in real time, which is where she saw what a digital double could do if you truly understood somebody, and she says privacy protection is what her current company fixed this time around. Her argument on this episode is that anomaly detection cannot answer the question that matters after a deepfake gets through the door, because an intruder in a stolen account touches what its owner would touch.
In this episode
| 00:41 | Welcome, and who Kris Bondi is |
| 01:24 | Everything good about AI is what makes it difficult in security |
| 02:06 | Why enterprises fall behind: a step behind becomes ten steps behind |
| 02:48 | The other side of the ledger, and whether AI helps the defenders |
| 03:00 | Red team, blue team, and the companies that cannot have either |
| 03:51 | The second benefit: reading signals and responding inside a process |
| 04:36 | Her own product as the example, and the composable model built on the fly |
| 05:43 | No, two different things, and anomaly detection on steroids |
| 06:23 | From something odd with the account to certainty it is somebody else |
| 06:59 | The next capability: querying an LLM about intent |
| 07:40 | A stolen password, and what AI can tell you without person-level matching |
| 08:08 | Switching to deepfakes |
| 08:26 | Deepfakes as an arms race that will keep escalating |
| 08:53 | Detection at the door: IDs, and the glitches in a face or a voice |
| 10:14 | Continuous validation, past the face and into how somebody interacts |
| 10:45 | What a successful deepfake leads to: fraud, ransomware, stolen data |
| 11:17 | The systems admin case, and touching what an admin would touch |
| 11:41 | So this is monitoring behavior for anomalies |
| 11:47 | Why the anomalies may not show up |
| 12:24 | The signals: mouse, keyboard, and the ones she will not name |
| 12:52 | Two technical co-founders who build Unix command lines differently |
| 13:24 | A fingerprint, and why she calls it multidimensional |
| 13:41 | The left pinky, and the quirk a deepfake cannot know |
| 13:52 | What inspired her to start the company |
| 13:58 | The luck society has had so far, and why the company exists |
| 14:17 | The Israeli neural network, the digital double, and the privacy fix |
| 15:06 | The Social Security change in that day’s paper |
| 15:39 | Why it goes beyond voice, and the clerk who knows neither caller |
| 16:26 | Voice cloning, and what the Social Security office is not doing |
| 17:06 | Her worry about moving everybody online instead |
| 17:22 | 200 million bots at a bank, and why that is not unheard of |
| 17:49 | Where does this fit: industry, use case, size of company? |
| 18:01 | Manufacturing, financial services, online services, cybersecurity |
| 18:21 | 250 to 15,000 people, and the CISOs, CIOs and CTOs who buy |
| 18:55 | The closing question: advice, and where to look next |
| 19:17 | Innovate your security the way you are innovating everything else |
| 19:41 | The question to ask any security tool: how adaptive and dynamic is it |
| 20:01 | Real time has to mean action, not one more alert |
| 20:22 | Close |
In Kris’s words
“everything good about AI is a part that makes it difficult in security”
Kris Bondi (01:24)
“they start a step behind the bad actors and then quickly could end up 10 steps behind because they haven’t evolved and they haven’t innovated”
Kris Bondi (02:06)
“I can tell you with certainty that’s not Christina using it. That’s David using it.”
Kris Bondi (06:23)
“the anomalies may not show up because they’re touching the things that they normally would touch”
Kris Bondi (11:47)
“There are just things that people do that are inherently specific to them.”
Kris Bondi (12:52)
“Anyone who’s building a deep fake of you is not going to know that you have this quirk of your left pinky caps on and caps off.”
Kris Bondi (13:41)
“I would hope that if they are listening to your podcast and are on a journey of implementing AI, that they extend that into how they are innovating their security as well.”
Kris Bondi (19:17)
“Is it operating in real time? And not just to send an alert, but actually to take action in real time.”
Kris Bondi (20:01)
Resources
Kris Bondi on LinkedIn: Her profile, where she posts on identity, deepfakes and enterprise security, and the source of her title as CEO and co-founder
Mimoto: The company she leads, which identifies people rather than credentials. Its own wording for the product is person-based continuous validation, the idea she describes in her own words as continuous validation
Ideas and terms discussed
Anomaly detection on steroids: Her phrase for what compute now makes possible in the standard control: analyzing what is happening inside systems, faster and wider than before. She uses it to mark the boundary of what that control can do, because it reports that an account is behaving oddly and cannot say who is behind it
Person-level identification: The half of her argument that anomaly detection cannot cover, and the reason she describes the two together. Advanced pattern matching identifies the individual from behavior rather than from credentials, so the alert changes from something odd with an account to certainty that the person using it is somebody else
Continuous validation: Her name for checking that it is still the right person after authentication, expanding beyond just the face into how somebody interacts. The company site uses the same words for the product
Advanced pattern matching: The technique underneath, which she is careful to say is not new. What is new is that it scales, which she credits to where ML and AI are today, and scale is what makes it fast enough to respond with rather than to review afterwards
A composable model built on the fly: How she describes her own system working: data streams come in, the algorithms are chosen according to which streams they are, and a model is assembled in the moment to make a determination and take action in real time. She dates her own product’s capability to about three years before the recording
Red team, blue team, purple team: The capability she says AI now puts within reach of companies that never had it. A red team tries to breach you, a blue team tries to stop it, and a purple team does a little of both, and the change is who can afford one rather than what one is
Querying an LLM about intent: The capability she is waiting for, and the only prediction she makes with a horizon on it. Rather than flagging that behavior is unusual, a system would be asked what it looks like somebody is trying to do, and she expected that potentially within the year at the time of recording
A multidimensional fingerprint: Her upgrade to the fingerprint metaphor Christina offers. The identifying signal is not one habit but many at once. Her own example is her left pinky on the caps key, which she then rules out as an identifier of herself, because she has said it in public too often; the point is that anyone building a deepfake of somebody else would not know their equivalent
The digital double: The idea she brought with her from an earlier company, where a neural network predicted what people were doing offline in real time. The insight was how far you can go if you truly understand somebody; privacy protection is what she says her current company fixed this time around
Account takeover: The mechanism that connects the two halves of the episode. A deepfake at the door often results in an account being taken over, and the damage is what the intruder touches afterwards rather than the entry itself
Named on air
ElevenLabs: Christina raises ElevenLabs as an example of how far voice cloning has come, including cloned voices carrying whole conversations that are indistinguishable from the real person
The Social Security Administration: Christina raises that day’s news that benefits enrollment and bank account changes could no longer be done over the phone. Kris Bondi’s answer moves the problem off voice cloning and onto the process
Frequently Asked Questions
-
Continuous identity validation is checking that the person using an account is still its owner after login, not only at the moment of login. Kris Bondi of Mimoto describes it as expanding beyond just the face, into how somebody is interacting, which she calls the other place to catch an intruder who has already gotten past authentication. The signals are behavioral: how a person uses a mouse, how they use a keyboard, what they do at a web level, and habits as specific as the way two engineers build their Unix command lines differently.
Transcript 10:14 to 13:41
-
Anomaly detection misses a stolen account when the intruder does what the account owner would normally do, because there is no anomaly to see. Kris Bondi of Mimoto puts the failure in one sentence: the anomalies may not show up, because they are touching the things that they normally would touch. Her example is a systems administrator account, where the person who broke in touches exactly what an admin touches, at hours an admin may also be working. What is needed instead is an answer to a different question, which is whether the account is being used by its owner.
Transcript 11:17 to 12:24
-
Companies defend against deepfakes in two places, at authentication and afterwards, and Kris Bondi of Mimoto calls the first one an arms race, in which attackers keep getting better and detection has to keep innovating. At the door, outside organizations specialize in detecting deepfakes, companies using government-issued IDs often run something else behind the scenes as well, and visual and audio systems look for glitches suggesting the face or voice is not the right person. Past the door, the second layer watches how the person interacts, on the assumption that something got through.
Transcript 08:26 to 10:45
-
After a deepfake gets past authentication the result is often an account takeover, and what follows depends on which account. Kris Bondi of Mimoto separates the anti-fraud side, where deepfakes can cost companies a lot of money, from the intrusion side, where a deepfake used to get into a system is where a ransomware attack may happen or data may be stolen for espionage. Her worked case is an internal takeover of a systems administrator account, where the damage is not the entry but everything the intruder touches afterwards.
Transcript 10:14 to 11:41
-
AI can be used to test a company’s own security, and Kris Bondi of Mimoto names it as the first benefit AI brings to defenders. A red team tries to breach an organization and a blue team tries to stop it, some companies run a purple team that does both, and many organizations do not have the ability to run either. Telling AI to figure out how to breach your systems can now be run in-house or bought from service companies, which she says puts the capability within reach of companies that could not have it before.
-
AI makes attackers more dangerous by letting them change approach, innovate on it and automate at scale, all faster than the organizations defending against them. Kris Bondi of Mimoto frames it as an asymmetry rather than a new weapon: everything good about AI is part of what makes it difficult in security. The second half of her argument is about pace, and it is the half enterprises can act on. Companies tend not to implement or innovate as quickly as attackers do, so an organization that starts a step behind can quickly end up ten steps behind.
Transcript 01:24 to 02:48
-
Executives should ask how adaptive and dynamic a security tool is, meaning whether it can respond to the problem as it looks today and as it looks tomorrow. Kris Bondi of Mimoto adds two tests underneath that question. The first is whether it operates in real time in the sense of taking action, rather than only sending an alert. The second is whether it enables the team to respond faster, or simply adds more of a tidal wave on top of them. Her related advice is to extend the innovation of an AI program into security as well, rather than using old security for new problems.
Transcript 19:17 to 20:22
-
Person-level security monitoring is aimed at mid-size enterprises, which Kris Bondi of Mimoto puts at roughly 250 to 15,000 people, though she says her company can work across everything. The industries she names are manufacturing, financial services including regional banks and credit unions, online services, and some cybersecurity companies that build it into what they deliver. The buyer is sometimes a CISO, sometimes a CIO or CTO, and highly regulated companies tend to want it. Her own test is not the org chart: the best fit is people who have had to clean up the mess, or who will have to.
Transcript 17:49 to 18:55
-
[00:41] Christina Ellwood: Welcome to AI Realized podcast for executives leading AI deployments. From addressing security, data, and operations challenges to managing the organizational changes, AI deployment re- represents the opportunity to redesign our organizations from the inside out. I’m Christina Ellwood, your host for today’s episode, and we’re talking today with Kris Bondi, the CEO of Mimoto. Welcome, Kris.
[01:07] Kris Bondi: Thank you very much for having me, Christina.
[01:10] Christina Ellwood: It’s delightful to have you with us. I really appreciate you joining us today. I know that you’re a security expert, and I wonder if you would explain to our executive listeners how AI exacerbates security problems.
[01:24] Kris Bondi: AI-- Uh, everything good about AI is a part that makes it difficult in security So that’s the easiest way to say it. So what that means is that AI enables, uh, bad actors to be able to change their approach to, to, if we’re saying breaches or malicious activity, to change their approach and to innovate on it, and to innovate and also have a evolution of it. It also enables automation at scale. So that means-
[02:01] Christina Ellwood: Okay, so we’ve put a superpower tool in their hands to make it easier for them to hack.
[02:06] Kris Bondi: Yes. And the other part of that is a part that companies often don’t think about, is that ability to continue to innovate and evolve. The-- a byproduct of that is that enterprises tend not to implement as quickly or innovate as quickly, and so staying... Sometimes with some organizations end up being they start a step behind the bad actors and then quickly could end up 10 steps behind because they haven’t evolved and they haven’t innovated in how they are, how they’re responding or how they are setting up new barriers.
[02:48] Christina Ellwood: It’s always been a bit of a ca- game of catch up to keep our organizations secure, so it sounds like we’re just, we’re running slower than the hackers are. But are there some benefits to AI for our security posture?
[03:00] Kris Bondi: Most definitely. There are several things. One of which is many organizations don’t have the ability to have a red team or blue team. So red team is an internal team that is trying to, trying, frankly, trying to breach you. Think of it like figuring out how you can be breached. And, and on the other side, a blue team is the team that is responsible for trying to stop that. Sometimes companies will have what you’ll hear a purple team, which is a little bit of both. That can now be done externally or that can be implemented within an organization where you have AI. I’m saying to AI, “Figure out how to breach my systems.” Or how could some-
[03:42] Christina Ellwood: Are these services that are offered?
[03:43] Kris Bondi: These are services that some-
[03:45] Christina Ellwood: That I could h- I could hire a service organization?
[03:46] Kris Bondi: There are organizations that do that or service companies that do that. Um- but the ability to bring that in-house or be able to have that makes it within reach for companies that couldn’t have that before. That’s one. The second thing is- Yeah. If we are, particularly for your team that has people who are implementing AI solutions, the ability to do that same analysis internally of not just from the where’s the breach, but how to move things along in a process so that there are less touchpoints or there may be spots along the way where you’re checking them, but the automation of, of being able to read signals and respond is something that is-- that wasn’t possible. What-- So I’ll give you an example from my own company of what we do wasn’t possible three years ago. We take a number of data streams, and depending on what data streams they are, we pick which algorithms are being used, and on the fly we’re creating a composable model to be able to make a determination and take action in real time. All those words I said, that would’ve been such-- those would’ve been pieces that you could potentially do. We do advanced pattern matching. Advanced pattern matching never scaled before. We can scale it because of where we are today with ML and AI, and that ability to be able to be faster in our response is something that companies now have that they didn’t have in the past.
[05:19] Christina Ellwood: So let me make sure I understood what you just said. So using advanced pattern matching, you are able to evaluate what is happening dynamically in the organization’s traffic or patterns of use and identify when there is a breach? Is that correct?
[05:43] Kris Bondi: No. No, there’s two different things. Sorry, I did say it really fast. Using-- So AI enables... AI and where we are today with compute enables us to be able to do everything you said about analyzing what’s happening within systems. That is a anomaly detection on steroids, is probably a good way of saying that. The advanced pattern matching that we do that we couple with that is we’re actually identifying a person based on a number of things, and the advanced pattern matching is saying it’s not just an anomaly of there may be something odd with Christina’s account. It is, “I can tell you with certainty that’s not Christina using it. That’s David using it.”
[06:28] Christina Ellwood: Oh, I see.
[06:29] Kris Bondi: And so that is the being able to identify who the person is based on because of what we’re doing of identifying individual people, and that’s an advanced pattern matching. But when you put those together, it gives an incredible, an incredible strength, and that whether you have our advanced pattern matching or not, where we are today with, with AI and ML enables companies to be able to do that, a detection of what’s happening and read things that they couldn’t read before. And also we’re, we are not quite there yet, but the expectation is that we are very close to having the point where an LLM can be queried to be able to ask intent. So I’m seeing that Christina is doing this and this, and not just saying, “That’s odd. Christina doesn’t usually do these things,” but having it be, “Christina’s doing these things. What does it look like she’s trying to do? Oh, it looks like she’s trying to map the network.” So we’re not quite at that point, but that’s where we’re going, and I would expect that potentially within this year of LLMs having to, having developed to the point where we can actually query it about an intent.
[07:40] Christina Ellwood: Okay. That, that’s really helpful. So essentially, even if someone were to steal my password and log in as me, if my company was using Mimoto, you would be able to identify it’s not me, and companies can use AI and ML to do similar kinds of things even without Mimoto. Is that right?
[08:05] Kris Bondi: Not at the person level, but at the anomaly detection level, yes.
[08:08] Christina Ellwood: Okay. All right, great. Thank you for separating that for me. That’s very helpful. So I wanna switch to talking a little bit about deepfakes, because that’s been such a big issue in the marketplace. Can you talk a little bit about how companies can safeguard from deepfakes?
[08:26] Kris Bondi: Yes. I will also start by saying I think deepfakes is a, is an arms race I think that going back to the bad actors innovating, they will continue to get better, and how to respond to deepfakes or how to catch them in the action will need to continue to innovate.
[08:47] Christina Ellwood: Maybe you’ll break those down. Yeah. The identify and the response. Maybe you’ll take, give both-- give us a little background on both of those.
[08:53] Kris Bondi: Yes. So the, so the ways that, again, these are outside... There’s outside systems, outside organizations that are specifically focused on identifying deepfakes when it comes to the authentication spot. So companies that are using government-based IDs often are not only using that, they often have something behind the scenes that they are, that they’re using. So there, there’s one version. There’s another version where it is there are companies that are doing either visual or audio or both where they are... And this is where I was saying it’s an arms race of looking for certain glitches, for lack of, in a layman’s terms, looking for certain glitches that are make you aware that may not be the right person or that the voice... That when, that it’s not the right person. Whether it is a person but using somebody else’s voice, which is a... So it may look like the right person, but it’s different words coming out. Or it may be, um, somebody representing themselves but, uh, or representing a version of themselves, but they’re being fed information from another place. All those are in the deepfake category. So that is the getting in the door part, and that’s the part where I was saying that’s truly is an arms race. On the other side of it, it is catching if somebody’s gotten past that, how do you respond to it? Being able to catch at a additional level, and this is the continuous validation of where it expands beyond just the face. So how are they interacting is the other place to catch it. The danger with... One of the big dangers with deepfakes is there’s both on the anti-fraud side, and that’s costing companies a lot of money potentially. On the other side is the, if they’re doing a deepfake to get into a system, that’s where the ransomware attack may happen. That’s where the, that’s where, from an espionage standpoint, data could be stolen. And the challenge with those is that often is an account takeover, a internal account takeover, which may mean that if I was a systems admin and there was an account takeover, so we’ve used a deepfake to get in. I’m a, I’m a systems admin. The things that I’m touching would end up being something that the person who is trying to breach, having gotten in the door with a deep fake, would touch. If I’m doing that at, say, midnight my time, an admin may be working at that time. And so you have people who are... It’s not just the danger of them getting in the door, but it is what they’re touching afterward.
[11:41] Christina Ellwood: So this gets back to monitoring the human behavior for looking for anomalies, right?
[11:47] Kris Bondi: It is. The challenge with that is the anomalies may not show up because they’re touching the things that they normally would touch. So it-
[11:55] Christina Ellwood: Would normally would touch.
[11:55] Kris Bondi: Mm-hmm ... they would normally touch, which is the reason why you have a solution like us that is looking at, but is it that person?
[12:03] Christina Ellwood: So you’re able to go, even if I am doing f- i-if my, my, my persona has been stolen and someone has used it to get in to the enterprise system, and they’re doing things I would normally do, you’re still able to figure out it’s not me.
[12:20] Kris Bondi: Yes. Because we’re c-
[12:22] Christina Ellwood: And how do you do that?
[12:24] Kris Bondi: Because we are looking at a person level, and so we’re looking at a number of things. We’re looking at things like if it is the biometric-related things, it would be things like ways that you use your mouse, ways that you use a keyboard. It may be other, and I don’t wanna give it all away here. Certain things that you’ll do at a web level. I have had... I’ll give you one example that is slightly deeper, is that I have two technical co-founders. One has a computer science degree, one ha- was self-taught at an early age. They build their c- Unix command lines differently. There are just things that people do that are inherently specific to them. And it is... So it’s not just the... There’s a... This is the reason why I talk about needing context as well, 'cause it’s not just what’s out of the system or what’s out of normal, but it is what is the thing that we always do. Mm-hmm. One of the things that-
[13:23] Christina Ellwood: It’s like a fingerprint.
[13:24] Kris Bondi: It is a multidimensional fingerprint. Something that I do that’s unique to me is my left pinky caps on and caps off constantly. That is... I’ve now said that enough times in public that we can never have that be an identifier of me. But if you were doing that, we would... Like, that would be something of Anyone who’s building a deep fake of you is not going to know that you have this quirk of your left pinky caps on and caps off.
[13:52] Christina Ellwood: Gotcha. Gotcha. Kris, tell me what inspired you to start Mimoto?
[13:58] Kris Bondi: I have a very firm belief that we as a society have been lucky to date that there hasn’t been a larger loss of life or a disruption in the way people live because of somebody getting into a system and not being seen, and either taking it over or holding it hostage or deleting it. That’s why we exist. What I have in my past that in- has gave me the ability to realize we could do this is, I had been with an Israeli company that had created a neural network to be able to predict what people were doing offline in real time, and to have the products interact with people. And so I, I had this understanding of what a digital double, if we really understood somebody, could do. That particular organization wasn’t necessarily in need of it being very privacy protecting. That’s something we’ve fixed this time around. But, but understanding if you have a digital double, if you truly understand somebody, how much further you can take that.
[14:57] Christina Ellwood: I see. That, that makes sense. Uh, and you were ahead of your time 'cause things have only gotten better and harder and whatever. I don’t know if you noticed in the paper today, but the Social Security Administration announced you can no longer enroll your be- for your benefits or change your bank account over the phone. You now have to go online or go into a Social Security office. I’m not sure how they envision that happening for people who are ill and bed, bedridden or disabled, but that practical issue aside, do you think that is a direct reflection of the enhanced security issues associated with cloning people’s voice?
[15:39] Kris Bondi: I think it goes beyond voice, but yes. I think it goes... It is... 'Cause it’s not, it most definitely, you wouldn’t even have to clone their voice. If I was working, if I was working at a... The question is, do you actually have to talk to a person or not? And even if you do You and I are both female. We both have a name that is a Christine, Christina name. The person working there doesn’t know who either of us are, so we wouldn’t even have to clone a voice in that case. It would really just be, it, it would really just be starting from the point of needing to change it. I do question that sending somebody in, the logistics and the, even if you could get people to go in.
[16:26] Christina Ellwood: And the alternative is you can go online, which of course that’s not an enhanced security approach. I don’t know what they’re trying to solve there, but I did think about the incredible accomplishments that have been achieved by companies like ElevenLabs in being able to clone voice and to incre- use a cloned voice to create whole conversations that are indistinguishable from the real person. So it, it did get me thinking about that, but I take your point that it’s, uh, as far as we know, the Social Security office is not recording our voice for the purpose of comparing it to the next time we call.
[17:06] Kris Bondi: I am concerned about the switching people to online of how ready are they for the amount of bots that are, and sophisticated bots that will come at them.
[17:18] Christina Ellwood: Maybe they need to be using Mimoto, huh?
[17:22] Kris Bondi: I’m just leaving it there, not mentioning it. But I do know that there are, there-- A bank will get 200 million, 200 million bots coming at them to-- if it is in a relatively small amount of time for setting up accounts or trying to do distribution. That is not a... That’s a large number, but it’s not an unheard of number.
[17:49] Christina Ellwood: In your work i-in the marketplace, is there a particular area that Mimoto is ideally suited for? A particular industry, use case, type of company, size of company?
[18:01] Kris Bondi: Yeah. So we, we tend to work... We can work across everything, but we tend to work in manufacturing, financial services, particularly regional banks and credit unions, online services, and, and then there’s some cybersecurity companies that are using us as part of what they’re delivering as well. Size of company, we tend to-- tends to be mid, mid-size enterprise, so you figure anywhere from 250 people to 15,000 people. And then as far as, you didn’t ask who we sell into, but we tend to... Although we sometimes sell into CISOs, sometimes it’s CIOs or CTOs. I tend to think of it as people who have had to clean up the mess or who will have to clean up the mess tend to be very good companies for us. So th-that also means that companies that are highly regulated tend to wanna have us in there.
[18:55] Christina Ellwood: So I think that helps our listeners to understand when they should look to you for being a solution for them. What advice do you have for our listeners as we think about wrapping our conversation here? What advice do you have, and maybe what resources would you point them to if they wanted to look further into how to increase or improve their security posture in their enterprise?
[19:17] Kris Bondi: I think internally, they need to look at how they’re innovating, and I would hope that if they are listening to your podcast and are on a journey of implementing AI, that they extend that into how they are innovating their security as well. They’re not trying to use old security for new problems. So that’s one significant one. And then I’ll come back to, I believe that in assessing any new security or even their current security, they need to ask how adaptive and dynamic is it? So in other words, how will it, it be able to respond to the problem of as it’s presented today and as it’s presented tomorrow? How quickly can they find it? Is it operating in real time? And not just to send an alert, but actually to take action in real time. Or enable their team to take action quicker, and that comes down to the third part of the contextual part of are they, are they just putting more of a tidal wave on top of their team, or are they actually enabling people to respond quicker?
[20:22] Christina Ellwood: That’s great, great advice for them to take away. Thank you so much. So Kris Bondi, CEO of Mimoto, thank you so much for talking with me today on “AI Realized.”
[20:33] Kris Bondi: Thank you so much.