Perhaps the better plan is to accept that AI is at the top of its hype cycle, and, like any new technology, there will be some limitations to ChatGPT-style AI, which in its raw state can be subject to issues like hallucinations. We knew this anyway, as the CEO of the company behind it explained: “ChatGPT is incredibly limited but good enough at some aspects to create a misleading impression of greatness. It's a mistake to be relying on it for anything important right now.” 

ChatGPT is just one form of AI 

But therein lies the problem: ChatGPT isn’t AI. It’s one form of it. It isn’t predictive analytics AI (Machine Learning), which can help you analyse historical data to offer insights about potential future outcomes. ChatGPT isn’t Computer Vision, which is now so advanced it allows machines to interpret visual data to the extent it’s how your smartphone recognises your face and how autonomous vehicles can see the road. And it’s certainly not the end point AI researchers want to get to of Artificial ‘General’ Intelligence, AGI, which would be a type of artificial intelligence that matches or even surpasses human capabilities across a wide range of cognitive tasks, as opposed to the narrow, constrained problem sets we tend to apply it to now. 

And while I enjoy playing with GenAI as much as anyone, and certainly see it as a great aid in some forms of business content creation, at no point did I see it as the basis for a way to predict interest and recommend products based on a user's browsing history or purchase patterns – or what I’d recommend to my clients to use for processing large amounts of data or for uncovering insights on of the performance of their business, or guiding decisions in areas from marketing strategies to inventory management. 

AI can deliver groundbreaking initiatives  

But I have (and do, every day) tell clients that they should be using AI to do just those things. In fact, much more: for better customer relationship management, for accurate detection of fraud in real-time, for content moderation at Internet scale and volume, as an ideal way to improve visibility across their supply chains, for sales forecasting, improved fault prediction and quality control in manufacturing – and much more. I have worked on several large AI projects around, for example, aspects like the human genome and medical tracking of Olympics athletes and I have a good sense of what’s IT industry hype and what’s actually real, useful, and reliable enough to look to build your next wave of innovation on. 

I know AI can deliver this. I know we’re helping clients do genuinely groundbreaking things with it. But I also know that it would be naive to completely ignore some of the issues surrounding AI such as data bias, lack of governance, proven use cases and so forth. 

It is far better to take a pragmatic view where you open yourself up to the possibilities but proceed with both caution and some help. That must start with working through the buzzwords and trying to understand what people mean, at least at a top level, by an LLM or a vector search or maybe even a Naive Bayes algorithm. But then, it is also important to bring in a trusted partner to help you move to the next stage to build an amazing new digital product, or to undergo a digital transformation with an existing digital product. 

Whether you’re in startup mode, you are already a scaleup with a new idea, or you're a corporate innovator looking to diversify with a new product – whatever the case, you don’t want to waste time learning on the job, and instead want to work with a small, focused team who can deliver exceptional results at the speed of modern digital business. 

Get real about AI by getting real with your data first 

Whatever happens or doesn’t happen to GenAI, as an enterprise CIO you are still going to want to be looking for tech that can learn and adapt from circumstance and so help you do the same. At the end of the day, hype cycle or not, AI is really the one tool in the toolbox that can continuously work with you to analyse data in the wild and in non-trivial amounts. This allows you to work together to find the good solutions, adapt them to improve success rates and better model the fast-changing world the data is trying to reflect. 

There’s a lot more to successful AI adoption for innovation, too, than signing up for a trial version of the latest Google AI helper: it is really important that you clean your data and align your approach with the ethics of what you are trying to do and what it might mean for data privacy, and so on. 

But the bottom line is to think less about the headlines and more about what advanced, non-deterministic programming (in other words AI) could do for your brand – and how you’d like to turn that vision into a reality.  For those looking to learn more about AI please download our free guide for starting with AI, it is available here.