A well-known cloud hyperscaler was planning to build a datacenter in our community. At first glance, it seemed a good thing. After all, I work in the cloud, right? This is the future, right? We’re transforming lives, right?

But as it went through the building approval process the community began to push back fiercely and it was ultimately withdrawn.

Why?

Its presence did not have much perceived value to the community. Due to state laws, it would not generate any local tax revenue. It threatened the water table of the surrounding wells supplying peoples’ homes and farms. There were concerns about noise, light pollution, aesthetics, and the destruction of displaced wetlands. The promise of local jobs was moot because roughly 10 people were needed to run the mammoth operation.

They made the right call and withdrew it in the end.

Now, I don’t for a nanosecond think the hyperscaler had any malice in mind. They even offered to make some contributions to the local school system in lieu of taxes, but in the end it just wasn’t a good fit for the area.

That’s ethics in motion in our quickly-evolving world of cloud and AI.

Poor ethical considerations cost them money, time, frustration.

AI is moving faster than YOU are

AI is presenting wonderful opportunity to the world of technology, but not without ethical considerations. It has moved beyond providing advanced information to autonomous decision-making with agentic frameworks. It’s even making decisions about what can make decisions, as agentic frameworks spin up and spider across the digital landscape.

This is both a power and an Achilles heel. If something can make good decisions, faster, and better for us, it can also make bad decisions faster than we can clean up before catastrophe.

That means you need to control it (governance) and monitor it (observability).

AI can help you be a better citizen of your community

AI is a tool. It can help you provide more value to your community or employer. That almost makes it an imperative. We have a responsibility to serve others well in the spirit of constant improvement and mutual benefit.

But what if AI is making poor decisions?

That same speed and autonomy that drives value now makes you a greater hindrance to others.

The wild beast of AI must be tamed with governance.

Just like we all must learn to exercise self-control in our financial, eating, and other habits to be successful, so must AI. It must be cared for and maintained.

Speaking at a recent conference, the thought “The difference between catastrophe and progress is not capability, it’s governance.” was our mantra.

But I’d add to that, to say that governance implies ethical considerations.

Governance Is Imperative

Given the statistical nature of models and the threat of hallucination in autonomous decision-making, governance is imperative.

Those mistakes are costly.

It’s more than a human in the loop. It’s envisioning the solution with not only positive outcomes in mind, but the potential bad ones as well. Where could it go wrong? What will cause it to fall off the rails?

I’m reminded of a popular doorbell camera platform that advertised a new pet finding feature for their product.

Seemed like a fantastic idea! After all, who wants to lose Fido and have to chase after them around the neighborhood when your phone might tell you before it’s a problem?

But it created a pervasive surveillance system, sparking privacy backlash. Exacerbating the issue you had to opt OUT of it, meaning your camera system was going to participate unless you caught it and shut it down. I understand for it to work, broad surveillance is needed, but SHOULD it? And how would people feel about that decision?

It seems counterintuitive, but ethics and governance are components of speed.

If you wreck or get pulled over on the way to grandma’s house, it doesn’t matter if you were driving 120MPH beforehand, you’re going to be later than if you’d driven at a manageable speed.

Crash badly enough, you may not be alive to go to her house again.

And I say that as someone who likes to be first, likes to win the race.

What happens when the purchasing agent you built makes bad decisions faster than ever before?

Governance Requires Observability

This is where observability comes into play. Observability is how we keep humans in the loop at scale.

Do you have the logging needed to track what your agents are doing? This of course applies to more than just AI features and should include your broader platform.

Observability is a great place to use AI for good. Much like banks use it to capture fraudulent activity, you can use it to capture AI run amok.

AI amplifies data risk

Yes, AI can expose sensitive data and content. Privacy continues to be a major concern, and 90% of companies report privacy as an issue with AI adoption – you are not alone.

Another issue AI amplifies via bad results is from bad data. Bad results continue to be a top-reported issue with AI adoption across enterprises.

How do you protect privacy and avoid bad results? Are you considering data security and privacy concerns from the onset of your solutioning? Are you adequately preparing data for AI usage?

Are you using AI to help address these problems?

AI can help secure your platform

We’ve all received letters in the mail about a data breach with our information.

AI can be used for good in these scenarios to identify the breach before it’s widespread, it doesn’t have to be the perpetrator.

When I first heard about the 27-year-old OpenBSD TCP/IP security risk that was found by Anthropic, I was thankful for AI and the good it did there. It found the flaws.

Flaws are everywhere, but we don’t have the data retention, processing, and statistical power in the space between our ears to synthesize it as fast as an AI agent can inform us.

AI as a force multiplier for good

The reality is that AI is a force multiplier, no different than the personal computer was in the 80’s. It is not the goal, it is a tool to reach your goal. If you’re NOT using AI responsibly, you’re robbing your clients and employer of the value you COULD be providing.

64% of employees say AI is enabling innovation. Are you among them?

Because as surely as you don’t use it, someone else WILL, outpacing you, or using it for some nefarious purpose like an agent swarm attack.

AI for good

The reality is that AI is here. It’s going to be used. The choice is up to you to use it for good.

Next time you’re undertaking an AI project, what story are you writing? Is it one that benefits those around you? Does it do more than improve a bottom line? Are you caring for your community, its data, and the impact your solution offers?