We use cookies to ensure our website operates correctly and to monitor visits to our site. This helps us to improve the way our website works, ensuring that users easily find what they are looking for. To allow us to keep doing this, click 'Accept All Cookies'. Alternatively, you can personalise your cookie settings.

Accept All Cookies Personalise settings

TechWatch Live - VRAI 

Q&A: Niall Campion, Managing Director for Product and Customers, VRAI

Q: What does VRAI do? And how can AI work with immersive technology?

A: We make data for virtual reality simulations.

In general, we are an interesting confluence of technology – VR, AI, cloud computing, machine learning, 5G. And AI is important as we use it to capture data from users, so once we go to the simulator, we can capture eye-tracking output data and other big datasets that are too complicated for a human to comprehend. For instance, a current project has 1.2 billion data points. So we make that meaningful data. It’s really about making big data sets understandable.

Q: In terms of immersive technology, what do you think is particularly disruptive or innovative about them? How might both immersive tech and AI transform defence and security?

A: As a training company, what we think immersive tech lets us do is reduce the cost of simulation training. So traditionally when you think of simulation training for fighter pilots or surgeons, this incorporates big expensive pieces of equipment. Commercial technology helps us bring that technology to anyone who needs it pretty much.

When simulators used to be the domain of specialists, we now feel we can bring this down some levels, so training crews to drive vehicles, for example. It’s about bringing the technology to other people, allowing them to train in a safer and more measurable way.

Q: In terms of those measurements, could you give me an example and how AI is used?

A: So a project we are working on with the RAF is about codifying and assessing ‘airmanship’ in pilots. This is defined as a non-technical skill required to operate an aircraft. It’s currently quite a subjective thing to measure; it’s an instructor sitting beside the students saying ‘your airmanship is good’ or ‘your airmanship isn’t good’. Whereas with our technology, we’re now capturing 24 million data points per user, and then we can start to say whether a student has or hasn’t got airmanship based on hitting the metrics.

Q: Looking ahead now, 5-10 years down the line, what do you say is the most interesting development within AI that might help you in your business?

A: Not just for us specifically, but off-the-shelf tools that are available for anyone to start using AI in their own businesses. So in 5-10 years’ time I would see a situation in which, when people are confronted with a large data set, there’s an algorithm that can allow you to analyse the data to give back more interesting results. In a broader sense, there’s the question of ethical AI. Rather than it just being a black box that the data goes into and then spits out results, as people become more literate with AI they want to understand what’s being done and why.

Q: In terms of AI itself, what sorts of challenges do we face in the coming years with the way AI technology is developing?

A: So it’s probably the inverse of that, which is if it starts to become a black box technology. In this case, increasingly decisions are being made in our lives based on data that is generated. If we don’t have insight on how that data is being used, then that’s a real challenge. It’s why as a company we have always got transparency with what we do with the data and everything we capture. And that’s why technologies around ‘explainable AI’ are definitely going to become more important.

Q: What are the most exciting changes happening in the training space, looking one or two years ahead?

A: There are two things I would definitely like to see. We are focussing a lot at the moment on biometrics which can provide a whole new set of data for us to look into. The other would be haptics. Making the simulator smaller, but more high fidelity would be a really interesting development.