A3 - Association for Advancing Automation’s Post

Transcript

We're just looking back in history and thinking about all of those moments that have made our industry special. What stands out to you is some really important moments in automation. When I look back on my career, I think maybe there's been quite a few moments, but certainly some of those that stick out have been times when we've really changed a certain industry. So after I did my PhD in multi agent reasoning systems, I joined Motorola at a time when. For those of us that remember, mobile networks actually were. Slightly different than they are today. If you walked well, maybe not walked so fast, but it certainly if you moved very fast within a network, you'd probably drop a call. And of course, they're dealing with the perennial challenge of a fixed resource, which is the radio frequency and maximising getting as many calls that make money on it as well. So at this time, every second one device would be releasing a certain number of measurement reports. Each of them would look at six nearby cells. So up to 48 bits of data coming from every mobile every second. So colossal amount of data they clearly wanted to apply. What was I guess some of the fledgling AI on this as well? So at this time when I joined, it was really to write, build our own AI, things like genetic algorithm using simulated annealing methods to really interrogate multivariate environments to optimize those radio frequency networks. So some of the first AI systems, I think though a lot of people probably are under the false impression that it is new when in reality it has been around for a really long time. And maybe it's just more popular now or maybe we just have more like consumer facing applications of it. What would you say? Why? Why is it blown up? So we've had narrow AI for a while. Yeah, narrow AI that really helps you deal with. Individual use cases, be able to preempt and predict capabilities, be able to have use cases that were really designed on one specific set of data. What we're now seeing is generative AI that really is allowing more bespoke capabilities driven to the individual who's using its need, but also to operate over multiple types of data and also for more broad use cases. As well and and of course, different interaction modes as well. So you might actually use natural language to query some of this to actually be able to get answers out over an iterative reinforced learning with your input as well to that model, let's talk about your role as chairperson on the A3 board. What are some of your goals for the AI tech and strategy board? Well, look, it's a great time to be an AI and certainly the space of industrial. AI as well. And of course, all of that's now being influenced very heavily by what's happening in the field of generative AI. And I think from an AI tech board, we're really looking at how can we help our members to navigate this disruption that's happening because of generative AI. How can we help them really to make the appropriate strategic direction and decisions by providing them the training, the enablement, the vision. That we see in our members of the board see coming up as well, as well as, you know, being able to create education and communities. And be able to listen to what they want as well, such that we might be able to provide some policy direction or influence as well. So you talked before about the challenges of scaling AI across different departments. So how do you ensure uniformity in the application of filters and processes? Yeah, look, this multiple ways you can apply and bring AI to your organization. Of course, you can either use tools that are already there. And already imbued with AI. But then you can also build your own transformative applications and capabilities as well. In which case you might be building your own copilot environment or you might actually be building or tuning a model yourself as well for your own environment. And then of course you know, being able to apply security in other capabilities to it so. You know, whenever you're looking at that, suddenly if you're looking at that second scenario of building your own transformative environment, you'll want to ensure that your data has the same sort of unified environment that you're able to bring all of your data sets together to be able to maximise how you exploit that data. And then of course, you'll also want to be able to make a consistent view from creating those models to actually applying the same consistent values in those models to be able to. Um, then operationalized and scale those models as well. And in those values, I'm really thinking about how you ensure the same sort of filters are applied in terms of eradicating bias and, and other other things that you don't really want to see in the usage of those models as well. It's a learning journey, but we know each day which harms we want to avoid in the AI and just being up to date with the best tool sets and frameworks that actually. Come embedded with those guidelines and frameworks for helping you deploy will help any, any organization trying to deploy AI. In what ways do you see AI transforming productivity beyond just efficiency gains? Yeah. So look, generative AI has started a paradigm shift, OK. So it started a paradigm shift on how we create and use applications and I think are referred to that little bit earlier, right. So it's moving from its autopilot stage where it was doing. Is narrow environments to a copilot stage where it can actually have a tremendous opportunity in terms of revolutionising just about everything we do OK and so with intelligent apps, especially when you're creating your own apps and your own transformation environments, you can enable that natural language environment. You can actually have constantly improving user experiences. So you learn from that user experience and then feed it back in and then you can either deliver. You can deliver quicker new assets and features to market as well. So kind of enables you to really rapidly build in what you learn into that environment. And one of those sites is that it now becomes important for us to be able to require that anyone is able to create those new environments. And so democratizing that capability and the access to that capability requires that we have low code and no code environments.