Head of Digital Matt Prudente shares his expert view.
Q: Tell us what you know about AI. Do you see it as a positive or negative?
It’s a funny one. I see lots of positives and a lot of potential. But I also get everyone’s concerns, although maybe people have been watching Terminator too much!
Like anything, there are new tools that come along. I view AI as a tool. My simplest explanation is the development of a screwdriver. A basic screwdriver is great but requires a lot of effort. You can take some of that effort out by using an electric screwdriver. That’s progress, isn’t it?
It’s how I view AI. We have a lot of repetitive tasks that can be handled by technology. We do it all the time. The Industrial Revolution obviously had a huge impact. And now, with coding and computer tasks, I think AI has the potential to be a great tool. It already is.
We’ve seen it problem solve and build websites. Platforms like ChatGPT can generate code for software like WordPress. If you want a particular function, you ask for it and it’ll create the code to drop into your website. AI handles those more mundane tasks, so you can focus on the bigger picture.
Take Wix, for example. You can create an entire e-commerce website. It provides content, imagery and a template of your choice. It knows what works and can even apply SEO principles to it.
And then there’s medicine and science. It’s helping come up with new drugs.
AI will make many developers nervous but it still needs someone at the helm who knows what they’re doing. I think the future will be AI carrying out complete problem solving. It needs data input – visual, language or figures – to produce the solution.
Q: Have you tested it out? Any thoughts on which platforms are performing best (ChatGPT, Bard, Code T5 etc)?
We’re at the stage where we’re figuring out what suits us. We’re already using ChatGPT but there have been some failings. We create hybrid apps and use a framework called ‘ionic’. So I tasked ChatGPT with producing a particular kind of ionic app. It came back with a series of code and told me what to do. As I read, I could see that it was wrong and outdated. There’s a limitation there which means it must be checked constantly.
Having said that, as developers, we’re meant to do code unit tests, to check something’s working correctly. We create our own code to check the code. AI could perform a task like that in a much simpler way. We want to use these tools to make the process less manual.
Co-pilot, a Microsoft AI assistant, is something I’m really interested in it. The benefits will be spending less time on laborious tasks and seeing greater output – quantity and quality – meaning a better product for the client. It’d also reduce the cost of a project. As a business, we’ll be able to deliver faster and save money.
Q: How might AI benefit/impact your role in digital?
There’s always a worry but in any role, we have to keep up. Companies pay for employees to attend training courses to improve their skills, or upskill to an area that’s needed. We should always find a place where we’re needed for work. It just adapts over time. My job changed from working as a printer’s apprentice at 16 years old, to where I am now, heading up digital for the Surgery. I’ve constantly adapted and changed, but there’s always been a job for me. We just don’t know what that’ll look like in the future.
A hundred years ago, lots of labourers had their livelihoods taken away but people still have work and there are many more of us now. AI will replace some jobs, just as it has done in the past, but other jobs will come into play because human interaction is still needed.
Q: Do you think we’ll get to a point where we can teach machines ethics, empathy and compassion?
Those are human traits based on love – different kinds of love. As humans, love is there but we don’t always follow it. We don’t always show empathy and compassion. It’s something we need to cultivate in ourselves. That makes it a challenge, for me, to say that a computer could ever be sentient because it’s feelings, not data.
With programming, humans set the rules but AI sets its own. Rules are ok but principles override rules. If you have principles, you can have two rules that cancel each other out. How will AI manage that, given that compassion and empathy feed in? Love overrides so many rules. Traffic laws, for example, are put in place to protect but what if a decision must be taken that doesn’t follow the rules, even though it’s the right and ethical course of action?
I’ve been trying to think of another hypothetical situation that might relate to this. Let’s say you have an air-conditioned room full of people, that needs to be kept secure. How would AI deal with failed air conditioning while still considering the security threat? Would it open the doors to let fresh air in? The data set might dictate ‘if the temperature reaches 38C, open the door’. Or will it watch people’s expressions, to understand feelings and emotions, and respond? Might it respond intelligently by putting people in place to act as security? How might they feel? A lot of facial recognition AI can read expressions, but subtle nuances often don’t tell the whole picture.
Q: How do you think AI and humans will co-exist?
It depends on humans! We have choice. AI and technology all come from binary but with humans, we have a choice. It’s power. AI needs power. Computers. A man-made brain to be able to do this stuff.
When we have a power cut, it’s a disaster. AI is just technology, again. When your power tool runs out of battery, we use it manually. As humans, we have a choice as to what we do with it.
When the Industrial Revolution took place, it produced much bigger things and faster. And not long after that, we saw world wars and what was produced there. It can be great but destructive things were also created. There’s always the chance that humans can misuse tools. If you have a hammer, you can build or destroy. It comes down to human choice.
Q: Any other thoughts on AI, now and in the future?
It was 20-30 years ago that most of us were getting emails for the first time. From there, we’re using AI. We don’t have to be in an office – we can work remotely. That’s all come from huge, chunky computers. The advancements have been huge and it’s growing at breakneck exponential speed. Especially when computers are then figuring stuff out for you. The potential is amazing.
The pandemic changed the way we worked. We all had to jump on the digital bandwagon, whether we liked it or not. I wonder if that accelerated the need for AI.