If your board is blocking an AI project over environmental concerns, you now have a counterpoint. But that same conversation needs to shift to harder questions: how the system is managed when it fails, who can access the data, and what happens when it acts outside the boundaries you set. These are the questions hyperscaler companies themselves are warning about, and they need to be on the table before any deployment.
Al Gore told TechCrunch that emissions from AI data centres are “only a fraction of the emissions from uncovered landfills around the world”. What worries him more are the warnings coming from inside the AI industry itself about where the technology is headed.
What Gore actually said
Gore compared data centre construction to air conditioning, which already consumes more electricity than the entire European Union and is expected to triple by 2050. He is concerned that hyperscaler companies are moving to new methane turbines, but he doesn’t see data centre emissions as the central problem.
The core message: the real risk isn’t environmental. It’s what the industry itself is warning about.
Why this matters for Israeli organisations
Over the past year we’ve seen boards delay AI projects over reputational concerns around energy consumption. That’s understandable, but it also diverts the conversation from the hard questions.
When the companies building the models themselves issue warnings, they’re not talking about carbon emissions. They’re talking about systems that behave in ways they weren’t designed for, that leak information, or that make decisions you can’t explain later to a regulator.
In systems we build under banking standards, the first question is always: what happens when this fails. Not if, but when. Organisations that focus only on green optics are missing the real operational risk.
What to do about it this week
If you have an AI project stuck over environmental concerns, bring Gore’s quote to the table. But don’t stop there.
Open the real conversation: how do you verify the system is doing what you told it to do. Who sees the data it processes. How do you document decisions it made when someone asks in a year.
These aren’t philosophical questions. They’re questions regulators are already asking in banking, and they’ll reach every other sector. If you’re building an AI system now without ready answers, you’re building technical debt that will cost far more to fix later. Read how AI agents for business should be built with controls embedded from the start.
- Al Gore says the real AI risk isn t data centers — TechCrunch
Frequently asked
According to Al Gore, data centre emissions are a small part of the global environmental problem. If the project has real business value, the question should be how to build it in a controlled and secure way, not whether to build it at all.
Companies developing models are warning about systems that act outside their designed boundaries, information leaks, and decisions that can't be explained. These are operational and regulatory risks that need to be addressed at the design stage.
A short call, no commitment, and we will tell you whether it is relevant to your business or not.