AI – AN INDUSTRY SNAPSHOT

Last month we held the first BCS Breakfast Club at Moorgate in London which was attended by a select group of investors, developers, architects, consultants, engineers and end users from the data centre sector. We chose the topic of AI and asked participants to share their current usage, challenges, plans and predictions for the future. The discussion gave us a snapshot of views and insights from across the industry.

We hope it helps address the skills shortage.

Almost all of the group were in agreement that AI, in some form, can be used to take on time consuming administrative tasks. The hope was that this would free up time for senior executives to train the next generation of data centre professionals to support clients and help address the skills shortage. For many of the organisations around the table, this was the aim for the short to medium term.

ChatGPT is being used regularly.

Interestingly, ChatGPT is being used by most of the group members, mainly for reports and research. One of the attendees from a leading investment house said that they are regularly using it for detailed investor reports and memos. They encourage their teams to use it by asking a question, embracing the response but then critically reviewing it.

Another explained that their organisation is currently using ChatGPT to automate some admin, for example, scope to service and any repeatable documents. These are then reviewed by senior staff as they have found that it’s easier to do this than draft from scratch. This review has now been built into their process.

Others in the group don’t see ChatGPT as AI but more of an enhancer and they use it to produce reports in just one week that would normally take six months. One organisation is currently mapping their whole business process to then evaluate where AI tools can help. This is being done by a separate team of specialists – not engineers who have different skill sets. But in the future will they be the same people?

There are concerns around its usage

Some members of the group expressed some concerns around security especially around client data and stressed that clear policies on its use are a must.  Those that didn’t have those in place were in the process of developing them and discussion was given to the need to keep reviewing these as AI continues to evolve.

The importance of thoroughly reviewing any information was underpinned by a recent example in the US where two lawyers and a law firm were fined $5,000 (£3,935) after fake citations generated by ChatGPT were submitted in a court filing. The judge P Kevin Castel said in a written opinion there was nothing “inherently improper” about using artificial intelligence for assisting in legal work, but lawyers had to ensure their filings were accurate.

Some organisations are building their own

Some of our participants were working for organisations that are developing their own AI products, investing heavily in data science and resources. This underlines the strong belief that it will create valuable efficiencies and by building their own they are using data they know and trust that can be enhanced by AI. It was noted that no-one was at the stage where they were using it to do the thinking for you, as yet.

It will change the skill sets needed.

Some interesting points were raised about whether this approach to AI will change the skill sets needed in the sector and even prevent junior staff from learning. For example, when you are starting out you might have to file the scope to service documents and have to read every one - that is how you learned the basics. ‘If AI gets rid of these ‘menial tasks’ – how will the new generation learn? Will they just learn to manage the AI systems? Do we risk getting rid of the building blocks of our industry? Is this a good thing?’

Whilst acknowledging these concerns the Group felt that as technology moves forward we have to go with it or get left behind – that maybe we are just changing the building blocks and this is an opportunity to become more intentional in what we train.

It's not necessarily the silver bullet

The general view was that if we can feed in the correct data we should be able to use AI to speed up cost models and also use AI to stress test schedules. However, is it feasible to input all the knowledge and nuances developed over a 20 plus year career - probably not!  Some of the architect attendees felt that design was also an area that might not be suited to AI as you should never put forward a design without understanding why every element is as it is, and AI won’t necessarily explain the why.

One participant had asked ChatGPT to design a datacentre to see what the outcome was and it put forward ‘an OK process.’  However, there were concerns that you can’t trust the data – it’s free and public. For that reason, some of our participants see AI as a big risk and their current approach is cautious as although they don’t want to hold people back and appreciate that AI can spark ideas - we are operating in mission critical scenarios. We could lose control.

Overall the consensus was that as AI is only in its infancy we have to check and review it but as it evolves, the level to which we check it will become less. We will start to trust based on experience and it is then that our jobs will be different. Automation in manufacturing was flagged as an example where we trust machines to churn out complex parts and dimension check them – only manually checking about 1 in 100.

Looking forward

Whilst it is too early to predict the growth and impact of AI, it was generally agreed that early adopters will be the winners and success will be driven by those that invest. In the near future consultants and suppliers will be expected to use and embrace it as they do other tools and it will become the norm.

In an ideal world, there should be standard industry AI developed, and to be useful, it needs buy in from all parties or we risk a VHS/Betamax debacle.

Finally, everyone agreed that the end goal is to use AI and reap the benefits because staff shortages are still a real issue.  This is a compelling need to be open to it and see if it can help alleviate the workload for existing staff to enable them to refocus their efforts on training and upskilling others.