Artificial Intelligence is the next evolution of video analytics, but instead of building algorithms around something that we’re looking at, we’re building systems that actually learn what’s happening, on the fly. Inherently, the quality will continue to get better over time, and that’s going to help us to drive better adoption with better quality.
Another thing that we have to do is make the integrations easier to implement. Today, it can be challenging to implement these, so that’s a major focus Milestone has, driving more partner community involvement toward our Software Development Kit (SDK) to streamline the processes of integrating and installation.
The third component that is going to help drive market adoption relates to the massive amounts of data that we’re starting to collect. When considering all the device sensors that will be connected to systems, it makes it difficult to make heads or tails of that data today.
Joel: Do you think the evolution of AI will drive users and integrators to look beyond the security capabilities of systems to see what additional intelligence can be gained?
Mike: Absolutely. I think we’re still in the infancy of this, and we’ve talked a lot about driving business intelligence, especially in the retail sector. We hear that discussed frequently. We hear it a lot in the gaming arenas, as well. To your point, it hasn’t been widely adopted yet, but AI makes that more accessible.
Joel: What are some of the technical factors still limiting AI technology?
Mike: Historically, we would’ve said compute [power] was the problem. It takes such a massive amount of compute to process all this data. But as we’ve heard from NVIDIA, compute’s not as much of a problem anymore. I think the bigger challenge we face in the overall architecture and infrastructure of these projects is, how do we secure it? It is a huge amount of data and, if it’s valuable to us, it’s probably valuable to somebody else, too, so how do we keep control over that? That’s probably the biggest hole that we see in most implementations: a lack of attention toward the cybersecurity elements.
Joel: With artificial intelligence in its infancy within the security industry now, what do you see as the timeline — how do you expect this technology to progress?
Mike: Even just this year, we’ve seen advances in GPU architectures and GPU offloads, which are reflected in the Milestone XProtect Smart Client release that’s happening now. I think we’re going to see faster adoption than what we may have anticipated previously. If we take analytics as an example again, it had extremely slow adoption over time but — as with most technologies — it hit a tipping point where it started to move very rapidly.
AI, in and of itself, will make technologies move faster, which is incredibly exciting and something I don’t think anybody really anticipated. It’s hard to predict precisely, but I would imagine in three years’ time we’ll be having a different conversation about how we apply these technologies in different ways.
Joel: What will it take for AI to move from being a “cool technology” to a trusted, life-saving, crime-solving tool?
Mike: As always, the application of the technology and seeing it in real life sparks the imagination. I come from a solutions-engineering background. It was always my job to take the customer’s business problems and figure out how we can solve them with the software, other tools, or a partner’s tools. While that’s interesting, it’s a one-time solution. It doesn’t hit the masses, so being able to show how it actually works more broadly in real life to solve complex problems is super important.
What we heard [at the MIPS event] from Sergeant O’Hare and the City of Hartford gives us a real application of the technology, how they’re able to leverage it today. What they found out is that there are many other things they can use it for that they didn’t even realize — and that’s what people need to hear.