The big, audacious goals of AI remain the creation of enterprise-wide solutions which enable you to be far more strategic or faster to market.
To that end, Datacom is an accelerator of innovation for our customers and partners. We’re likely a bit different in our approach, as there’s a lot of high-level advisory around AI in terms of what’s possible. We take that to the frontline, from a practical perspective from the actual people who do the work.
Where are the top opportunities in the market for this type of AI?
JW: The possibilities for text-to-video AI across all industries are very impressive, when you consider the opportunities these advancements offer in terms of training or instructional support.
The video realism achieved by the likes of OpenAI’s Sora is quite amazing, especially when you look back to where this technology was just a year ago – it’s come a long way fast.
Consider the impact for medical, engineering or construction professionals, where procedures can be explained and instantly augmented into visual reality, when combined with technology like Microsoft’s HoloLens.
Similarly, educators would have the ability to bring their teachings to life, transporting their classes back in history, or elevating remote learning. For students, learning can be a far more personalised experience with the ability to record and summarise lectures through your own personal education assistant.
Although undoubtedly impressive, the technology still has a way to go in smoothing out the glitches. At that point, it’s a serious consideration for businesses to build a digital twin of its unique environment, recreating the real world in a context that makes sense for employees and customers.
What are some of the challenges this represents in the market?
JW: It certainly escalates the risks of deep fake videos being used to spread misinformation. I don’t think we’ll ever have another election where opinions are not influenced by automated bot farms and disseminated through fake social media accounts.
The world has already changed in that respect, as the hyper personalisation of search engines and automated marketing is just an everyday part of our lives.
That’s a challenge that all AI vendors are tackling and regulators may eventually require them to include watermarking technology in their AI-generated imagery that allows the public to tell whether AI is at the wheel.
For the incredible opportunities that we’ve spotlighted though, I see the potential of this technology as much more exciting than scary.
What are the complexities around managing these issues?
JW: More than ever, you need to adopt a position of critical thinking and zero trust to scrutinise exactly where information is coming from. This is increasingly complicated to do in practice.
We don’t see a lot of misinformation from the likes of hallucinating chatbots anymore, and bias has also been reduced. Thankfully, when it comes to factual inaccuracies, bias and use of copyrighted information, the major LLM developers have put a lot of guardrails in place.
As AI assistants become commonplace work tools, ensuring data security and privacy becomes paramount.
The rise of enterprise AI, with targeted models for specific industries and use cases, will not only unlock the full potential of AI for businesses but also provide the necessary controls to guarantee information democratisation without compromising security.
What are customers looking for at the moment when it comes to AI deployments?
JW: We’re helping our customers and partners to start unpacking these AI use cases by giving them ROI (return on investment) frameworks and helping them understand the data that fuels AI, including how to manage it effectively and securely.
There’s so much more data in play now, filling data lakes and warehouses. While our clients are modelling their own data for their unique business needs, they are also utilising the expanding ‘model gardens’, which enable engineers to start with a pre-trained model.
You may then train the solution for your own specific purpose, but generally customers are going to be chaining these models to build a multi-functional solution that can incorporate emerging tech like text-to-video.
We’re exploring these use cases with our customers, bringing them into a commercial perspective and at the same time, providing the human-led expertise that ensures safety and ethical considerations are maintained throughout every development stage.
Efficiency gains are the inherent goals driving AI considerations, whether that’s with the automation of regular or repetitive tasks or accelerating new initiatives beyond iteration and into development.
As our customers experiment with the technology, deploying it where it works well, we are supporting that process by connecting them to productivity-enhancing digital tools.
Just as we advise our customers who are exploring and deploying AI though, we approach the technology with our eyes open, guided by industry best practice AI ethics and governance frameworks. We also consider AI through the ROI lens.
Looking ahead to the near future needs of our customers, I think the next biggest challenge is going to be the democratisation of big data and the development of knowledge management systems.
Those systems will manage information from an array of systems across multiple disciplines and will require a unified search capability.
Those types of enterprise-wide services are a looming challenge, to securely enable access for all, to provide accurate and up-to-date information, and to make the processing of these services cost effective. It is going to be a fascinating journey – one that we’re looking forward to guiding our customers through.