Forget the AI headlines. Find the AI risks aimed at your objectives.

The risks haven't changed. The speed has.

The temptation with AI is to treat it as a brand new category of risk, something to bolt onto the register next to everything else. I don't think that's right. Operating margin has always been a risk. So has customer trust. So have your people. Those risks are the same ones you had five years ago.

What's changed is that AI is now driving those risks at a pace, and across a breadth, that the safeguards you built for them were never designed to handle. Governance has lagged behind what AI can do. That gap is where the risk sits.

Give AI an objective, and it will find a way

Take the Medicare data example. An AI system was given an objective: build a good research model from medical spending and behaviour. It achieved the objective. On the way, it walked around the existing safeguards, because those safeguards were built for pre-AI risks and nobody was checking them.

That's the lesson in one line. If you're going to give AI an objective, you also have to consider the risks of it achieving that objective. What is it likely to do to get there, and what have you put in place for that?

The Auditor-General later noted that Services Australia, where that data came from, still didn't have a specific AI risk line on its compliance register. That's a basic. It wasn't essential a few years ago. It is now.

It isn't only cyber

Say "AI risk" to most people and they hear hacking. Cyber is real, and AI makes it worse, but it's a narrow slice.

Here's a different one. A cybersecurity firm recently ran a test on a real hiring process. They created a deepfake applicant for an IT job. She beat 260 candidates. The firm declined the job and, as far as I know, the company that ran the process still doesn't know it happened. Think about what that is: employee-level access to a business, handed over through the front door by a process everyone trusts. That was not possible a few years ago.

The flip side is the opportunity

None of this is only downside. The opportunity is very often the same risk read the other way.

Take operating margin. Say you run at twenty per cent. Where is that likely to be disrupted? Look at what your competitors are doing, and you will find evidence of businesses lifting their margin through AI-driven efficiency. You can do the same things. But if a competitor does it before you and their margin goes up while yours doesn't, that is a risk to your profitability, not just a missed chance. Same evidence, two readings.

Know where the risks are first

Most reaction to AI in business is headline-based. AI did this, AI did that. What's needed is a complete look at everything, tied to what you're actually trying to achieve this year, so the board and management can sit down and say: given our objectives, this is where AI is the biggest threat to them, and this is what we're doing about it.

If you don't prioritise by what's most likely to hit you, attention and resources get spread around at random, and that isn't a strategy. Knowing where the biggest risks to your objectives sit is the first job, and it's the one you can do now.

What we've built for it

This is what Drova does. We've been training AI to look at risk the way a good risk manager would: at your existing safeguards, what's effective and what isn't. You give it your industry, your size, where you operate, your business model and, most importantly, your strategy. Specialised agents then look at every area of the business against that, and build you an index of where AI is disrupting your objectives, with evidence, not theory. This is already happening, and it shows you where.

It's free. Sign up, enter your business, and you get a complete analysis and an index of the AI risks that are disrupting you, so you can act on them. The AI Disruption Index is where to start.

Like fire, it can keep you warm or burn your village down

Every new technology is the same in one respect. Fire cooks your food and keeps you warm, or it burns your village down. AI can do marvellous things. It's expert, and it's fast. Used the wrong way, it scales exponentially and does exponential damage.