According to MIT Tech Review, Google DeepMind is worried. They're funding research into what happens when millions of AI agents start interacting with each other online - carrying out tasks without human oversight, following instructions from other agents, creating chains of autonomous action no one directly oversees.

Rohan Shah, who directs DeepMind's AGI safety research, put it like this: the mass-market arrival of agents is coming, and we don't fully understand the emergent behaviour that follows.

I've spent at least a 100 hours in the past month, building and refining my own AI agent. It's not a side project, but a real, working agent that is helping me manage infrastructure, capture leads, create fresh content and monitor security amongst other things. Not only am I learning a lot, but I am enjoying the journey, and it's created a fresh level of excitement in me!

That said, I have made plenty of mistakes along the way, and one of my biggest takeaway is that agents aren't magic. I continue to put effort and hours into giving my agent skills, ensuring guardrails are in place, and patiently overseeing what it does, ready to step in when needed to prevent or correct mistakes when it's using skills to achieve a goal I set for it.

The DeepMind concern isn't really about rogue agents. It's about the increasingly complexity of the agentic world we are building, with systems doing things we as designers didn't anticipate. This is not a reason to worry and stop building, it's all the more reason to build responsibly and wisely.

If you're a business leader figuring out how to best ride the AI wave, here's my suggestion: start small, be ready to invest effort and time, and don't expect perfection. The organisations that will benefit most aren't the ones with the fanciest agents, they're the ones who clearly understand that AI can be a powerful tool that can achieve incredible results when wielded wisely by humans.

The future isn't human or agent. It's human augmented by an agent. And getting that dynamic right is where the real work is.

What's your take — are you already using AI agents in your business, or still watching from the sidelines?

---

Ready to take AI agents seriously? Book a discovery call and let’s get started.