‘Just because it can, doesn’t mean it should’: How to use AI efficiently

5 days ago

Image: © Vitalii Vodolazskyi/Stock.adobe.com

Red Hat’s Jamie Hackett on how organisations can utilise AI properly through strategic planning and appropriate deployment.

“AI isn’t new,” says Red Hat’s Jamie Hackett. “It has been around since the early 1950s, but it was very much viewed as a scientific endeavour – it had limited practical applications in the business world – and was generally an opportunity reserved for large tech or scientific companies until recently.”

Hackett is of course referring to the explosion of AI interest and development sparked by the release of generative AI models such as ChatGPT – now a household name. As noted by Hackett, this AI revolution has led to all sorts of companies taking an interest in utilising the tech, either internally or as part of their software offering.

“We’ve gone from AI almost being a ‘nice to have’ to a ‘need to have’ while developing software.”

Hackett is a cloud consultant at Red Hat, where he helps customers utilise AI technology. He joined the company in 2020 as a graduate solution architect after completing a technology management degree at the National College of Ireland.

“My final year project focused on using machine learning to predict the stock market,” he says. “I still work for a living so it wasn’t as predictable as I had hoped!”

Resistance to change

With AI becoming as popular as it continues to be, Hackett recognises the dual presence of excitement and fear in people’s minds – particularly in the workforce, as worries of AI replacing human jobs remain frequent. However, he thinks that this is a “perception issue” and that ultimately, people are “frightened of the unknown”, drawing comparisons between the emergence of AI and other innovations such as the introduction of computer text editors, which replaced the need for typewriting skills in office settings.

The understandable AI-driven uneasiness experienced by the workforce is something Hackett says companies will need to overturn through honesty and openness if they want to introduce AI tools.

“A lot of resistance can come from not knowing the purpose of an AI tool or system that is being brought in. This can lead to people making negative assumptions and slow down uptake in the tool or system,” he says.

“If leadership teams can present AI as a force for good within an organisation and try to bring people along on the journey, it can lead to buy-in from teams. Organisations need to approach AI tooling as something that can improve their employees’ quality of life and not just something that can affect the bottom line.”

The little AI that should

When it comes to the actual integration of AI tools, Hackett says there are two major pitfalls that he sees organisations make; unnecessary deployment and generic tools.

For the first issue, Hackett sums it up simply; “Just because AI can do it, doesn’t mean it should.

“I’ve seen AI be used to solve a number of problems over the last few years and it reminds me of the saying ‘using a sledgehammer to crack a nut’. Many companies get hyped over AI and want to show progress so they use the technology to deal with a matter that could have been easily solved by just doing traditional software development.”

He says that he also sees a lot of companies buying and deploying AI tools that can do “generic” jobs, but are lacklustre when it comes to specialised tasks. “Most organisations struggle to list solid use cases that can deliver value in their organisations and as a result they go to the 101 of AI deployment and build something generic like an LLM-based internal chatbot.

“While it’s great to see organisations try AI and invest in the technology, generic approaches won’t return the value that you need to show to a board.”

Hackett suggests that companies should look to AI for fixing a specific problem that is too complicated to solve with traditional tech capabilities, or perhaps use AI as a second pair of eyes for peer review in the software development process, as he has witnessed a number of times.

He says that companies should always start with a small idea and deploy a cross-functional team to work with it and think strategically. “An AI tool that delivers immense and immediate value to even a small group of people can often deliver a lot more value than an overarching generic tool that tries to be a jack of all trades.”

Be responsible

As well as the strategic considerations that a company must keep in mind when venturing into AI territory, Hackett emphasises the importance of using the tech responsibly and ethically.

“A quick gut check is necessary from the get-go,” he says. “Would you like your personal data used in this manner? If it makes you uncomfortable then you probably need to examine your AI idea or project with a bit more rigour.

“A more formal approach would be to develop an internal organisation-wide strategy addressing the acceptable use cases of AI. You can use legislation such as the EU AI Act as a good baseline.”

He also suggests the use of tools that can detect and mitigate bias in AI models, such as open-source projects like TrustyAI.

At the end of the day, Hackett notes that AI isn’t a solution for every problem, but if used correctly and appropriately it can make a huge difference.

“Start small and ask the right questions; is this something we need to do with AI or can we do this with something else? Build a scalable framework for developing AI internally within your organisation, and make sure the controls are in place so that the data being used for AI is acceptable and ethical.”

Find out how emerging tech trends are transforming tomorrow with our new podcast, Future Human: The Series. Listen now on Spotify, on Apple or wherever you get your podcasts.

Colin Ryan is a copywriter/copyeditor at Silicon Republic