The collision of AI and security

The weekly newsletter of the Business of Tech, giving you new insights into the world of IT service delivery.

Looking for stories from the podcast stories? Check out the pod itself on Apple Podcasts, Spotify, or daily in your inbox. Stories are available to everyone for five days,and Patreon supporters forever.

Was this forwarded to you? Join the list!

The collision of AI and security

At the Business of Tech, I’m always on the lookout for the ‘why we do we care’ of any given topic. I’ve always operated with this principle in mind, but it’s been more relevant than ever as we navigate the many, many conversations floating around the AI stratosphere.

From out-there theoreticals to doomsday projections to best-case scenarios (that I often find too naive for comfort), I’ve been attempting to stay grounded with that same query: why do we care?

Last week, we answered this question with the Director of the MIT Center for Transportation and Logistics at MIT Dr. Yossi Sheffi. This week, we’re continuing along this down-to-earth path with an application-focused Business of Tech bonus episode conversation with Saagar Govil, CEO of industrial solutions and intelligent security systems provider Cemtrex Inc.

I really enjoyed getting in the headspace of someone who’s thinking about applying AI solutions both right now, and in the near future. Plus, he’s in the thick of the security space, which seems to be the second most important technology topic in our industry right now.

Keep reading for another real-world take on tangible AI.

The Overlap of AI and Security

Before AI took over the scene a few months ago, security was perhaps my most requested subject. So to kick things off, I asked Govil where he sees AI and security coming together.

Govil immediately confirmed my suspicions: we’re at a tipping point for AI’s utility in the security space. While AI has been around in the field for some time, it wasn’t really doing much for the industry from an analytics standpoint. But now, he’s excited about what’s to come.

In the present moment, he has his eye on AI’s ability to drive actionable events based on intelligent data coming in from security cameras. Though they’ve been able to detect motion and disruptions for a while, he thinks their analytic abilities are about to evolve: imagine detecting people and not cars, humans and not animals, and intruders crossing a fence, not just a person who could be a potential intruder.

In the future, he’s excited about ChatGPT powering some sort of real-time co-pilot to navigate a security breach – where the minute something happens, an AI can deploy an appropriate and nuanced threat response.

Wondering why these aren’t already the norm? The holdup, in Govil’s opinion, was on ease of use. He’s found that though customers have wanted this type of development, it hasn’t been handed to them on a simple enough silver platter.

Contextualizing AI as a Service Provider

Though Govil isn’t exactly in the MSP space, I was curious how he thinks these AI capacities should be labeled. Is it a tool? A feature? Something that’s getting layered in? A new product to invest in separately? If we want to be the people making money when our customers turn to AI, it’s worth analyzing the best messaging.

Govil’s response was simple: customers need solutions, and the practical deployment needs to be outlined for them. In his words:

“In the recent past, manufacturers would provide an AI technology that allows you to do analytics… a customer would say, okay, I want analytics, this sounds great. They would check a box, they would order the product, but then what would happen is to set up the analytics, it would take hours, or maybe even the person who’s responsible for installing it realizes the accuracy isn’t there, or there are too many false alarms. So this becomes a nuisance.”

The question he wants you to ask yourself is this: how do you deliver an experience that’s natively easy to deploy out of the box and that actually works from the beginning of the deployment? In plain terms, they have to see how it’s making their lives easier from the get-go.

An example he provided is one we’re all familiar with: Apple’s indexing of your camera roll to add names to each photo. They didn’t hand us the ability to tag people in the camera roll – they went ahead and did it for us.

To Govil, the key is simple and immediate applications – ideally in the background.

Beyond the Buzzword

By now, we can all agree that vendors are slapping on the AI label as a marketing tactic. How can we go beyond the buzzword to understand what real AI applications look like?

I asked Govil what questions he’s asking himself to navigate this challenge, and he started by noting that you have to be able to distill an AI’s function into smaller pieces, then unpack the nuances of each. Then, you have to ask yourself queries like…

Where’s an AI deployed– is it being deployed on the cloud? The edge? On-prem on a server?
What are the implications of that location from a bandwidth perspective? From a latency response time perspective?
What is the upfront cost versus the operating cost?
How is that data being used? Is it getting stored in the cloud? Is that a secure environment? Can it be hacked?

He also added this:
“And I think from an MSP or an integrator [perspective], you have to think about those kinds of risk profiles in terms of what your customer’s trying to accomplish, and then how to navigate that. But I think there’s a solution for everybody, and it’s just a matter of slicing and dicing what those different priorities are.”

Emphasizing the Data

The market is flooded with a gazillion different tools, and though we’re lacking objective industry standards right now, certain vendors will inevitably rise above the rest. Govil very much believes that the driving factor in who wins and loses will be data sets.

And not just the quantity – the source of data, and how that data impacts the model.

He predicts that because everyone is currently converging on the same sites to build data classification sets, we’ll reach the point where everyone can deliver similar results. At the same time, some companies like Reddit and Quora are either considering or have already blocked Open AI from crawling all of their data.

So, vendors with exclusive access to data sets will be able to train stronger models and deliver better products. He specifically thinks that vendors who learn how to make the most of their own unique data sets will reap the benefits here. For example, tying this back to the security camera example from up top, when Amazon with Ring and Google with Net jump on the opportunity to leverage their own security footage data sets, they’ll be the ones with the best AI model to pull from for smarter image interpretation.

If you’re already at the stage of recommending or shopping for vendors on behalf of your clients, this prediction is something to keep in the back pocket.

More info on Govil and his security work can be found www.vicon-security.com.

As usual, I want to hear where you’re at with AI, the MSP’s relationship with the tool, and its potential impact on business. Send those thoughts my way at [email protected]! We’ll be back next week with another industry deep dive.

More from MSP Radio

Missed Things?

How about our latest videos to catch you up?

The Daily Podcast available as videos

Building high bandwidth remote teams and how AI helps with Andy Tryba

Marketplaces, Influencers, Pax8 and SMBs with Shirish Nadkarni

Strategy in the Digital Age: Digital Transformation made practical with Michael Lenox

Limiting Hallucinations, Category Theory, and predicting policy with Eric Daimler

Want the Daily News?

All the stories from the daily Business of Tech Podcast are available in the daily digest, and stories are available to everyone for the first five days, and Patreon supporters forever. Catch the audio of the show anytime on Apple Podcasts, Spotify, YouTube, or wherever you find podcasts. Links at businessof.tech

Want to change how you receive these emails?
You can update your preferences or unsubscribe from this list.