‘Age of AI’? Not before the ‘age of data resilience’

Nearly two decades ago, Clive Humby quite accurately predicted that data would become the most valuable resource in our economy, coining the infamous phrase “data is the new oil”. Now, with artificial intelligence (AI), we’ve got the new Apollo 11. At times it can seem like the AI hype has reached its peak, but in reality, it’s just the latest chapter in a long-running story - digital transformation.

But the AI hype surrounding every industry at the moment is valid. It has immense potential to completely revolutionise pretty much everything, but we need to take a beat before we get ahead of ourselves. Organizations need to implement processes that can power data resilience and ensure that their data is available, accurate, protected, and intelligent to ensure their businesses can continue to run no matter what. 

Stay out of the shadows 

It’s far easier to manage AI adoption now with additional training and controls early on when it comes to something as all-encompassing as a company’s data. You don’t want to be left trying to untangle a mess later down the line. The best time to start was yesterday, but the second best time is now. According to the latest McKinsey Global Survey on AI, around 65% of respondents stated that their organization regularly uses Gen AI (double from just ten months before). But that’s not the scary statistic. Nearly half of the respondents said that they are ‘heavily customizing’ or developing their own models. 

This is a new phenomenon known as ‘shadow IT’ – unknown or unsanctioned use of software, or systems in an organization. For a large enterprise, keeping track of the sanctioned tools in use across various business units might already a challenge in itself. Individuals or even departments building or adapting large language models (LLMs) will make it nearly impossible to manage and track data movement and risk across the organization. But, putting processes and training in place around data stewardship, data privacy, and IP will help. Even if they help in no other way, having measures like these in place makes the company’s position far more defendable should anything go awry.  

Keeping an eye on things

Now, this doesn’t mean you need to put a ban on AI as a whole. It’s a great tool that organizations and departments will get significant value out of. But as it becomes part of the tech stack, we need to ensure that it falls within the rest of the business's data governance and protection principles. For the majority of AI tools, it's about mitigating the operational risk of the data that flows through them. Generally, there are three main risk factors: security (what if an outside party accesses or steals the data?), availability (what if we lose access to the data, even temporarily?), and accuracy (what if what we’re working from is wrong?).     

It’s in these cases where data resilience is crucial. As AI tools become embedded in your tech stack, you need to ensure visibility, governance, and protection across your entire ‘data landscape’. Uncontrolled use of AI models across a business could create gaps. Data resilience is already a priority in most areas of an organization, and LLMs and other AI tools need to make it a priority just the same. You need to understand your business-critical data and where it lives across your entire business. Companies may have good data governance and resilience now, but if adequate training isn’t put in place, uncontrolled use of AI could cause issues. The worst part is, you might not even know about them.

Strengthening your data resilience 

Data resilience is no mean feat - it covers the whole organization, so the whole team needs to be responsible. It’s also not just a one-time thing, data is constantly moving and changing – and so should your data resilience practices. AI is just one example of things that need to be reacted to and adapted to. Data resilience covers everything from identity management, device and network security, and data protection principles like backup and recovery. It’s an intensive de-risking project, but for it to be effective it requires two key things: visibility, and senior buy-in. Data resilience starts from the top down. Without it, projects fall flat, funding limits how much can be done, and protection/availability gaps appear. Data resilience is everyone’s problem.

But don’t let the size of the task stop you from starting. It’s impossible to solve everything all at once, but you can start somewhere. Get ahead of the game as much as possible before LLMs have appeared across your organization. It’s likely that many companies will fall into the same problems at they did with cloud migration, going all-in on the newest tech only to wish they’d thought it through a bit more at the start. Start off small(ish), with drills to test your resilience. After all, you can’t learn to ride a bike from a book – the only way to learn is to do. And don’t hold back, make sure you’re testing realistic worst-case scenarios. Try one without your disaster lead for example, a real emergency won’t wait around for them to be in after all. After running these tests. You’ll be able to gauge accurately just how prepped you are. Then, just start – if you don’t, you’ll never improve.