Switch color mode
Home Blog Can AI and Sustainability Co-Exist? Predictably, the answer is – it depends
The sustainability debate around AI is not going away, and frankly, it shouldn’t.
Andy Bailey
April 28, 2026
I’ve sat in enough boardrooms and marketing strategy sessions to recognize the pattern: when an entire industry starts brushing off a criticism, it’s usually because it hit home… The environmental cost of artificial intelligence is real, it’s growing, and any company that claims otherwise is either uninformed or being deliberately evasive.
But I also think the conversation has gotten stuck in a way that’s becoming its own kind of problem.
At Be My Eyes, we’ve had to confront a question that I suspect most AI-adjacent companies are quietly grappling with…
Does the impact of this technology warrant its carbon footprint?
Our answer is yes. But it comes with conditions. Understanding those conditions, I’d argue, is the more mature and honest way to think about AI’s place in a world where environmental accountability is no longer optional.
AI is not frictionless. It’s not immaterial. It does not live in some weightless digital ether.
Training large models requires enormous computational power, and consequently, AI data centers consume significant amounts of electricity and water for running and cooling that power. The hardware running all of this depends on complex global supply chains, critical mineral extraction, and manufacturing processes that carry their own environmental burdens. And, as AI becomes embedded in more products and services, these aggregate costs are only heading in one direction.
The sustainability debate around AI is not a PR problem to be managed. It’s a legitimate accountability question and one that deserves transparency from the companies building and deploying these systems.
Trust starts with honesty.
The responsible path is acknowledgement, not deflection. And that’s where I’d encourage every executive in this space to start.
Here’s where I want to push back on how this debate usually gets framed.
The dominant narrative positions AI and sustainability as opposing forces. As if responsible technology leadership means choosing one or the other. I understand why that framing exists, but it’s often too simple to be useful.
Nearly every transformative technology has carried tradeoffs. What determines whether those tradeoffs are justified isn’t just the scale of the cost; it’s whether the value being created is meaningful enough to warrant it, and whether we’re actively working to reduce the cost over time.
The two questions that belong together are…
What does it cost?
And what benefit to humanity does it really bring?
An AI system that automates trivial engagement, optimizes marginally better ad targeting, or shaves seconds off a workflow that already worked fine, or worse still generates unrealistic social media images just for fun—is probably not the most efficient application of AI.
But an AI system that describes to someone who is blind or has low vision where their Uber is waiting, or which of these ten cans on a supermarket shelf are the canned tomatoes, is something categorically different. It can, in some cases, be life-changing.
Be My Eyes exists to make the world more accessible for people who are blind or have low vision. More than 340 million people globally are blind or have significantly low vision, and that number is projected to grow significantly in the coming years. For many in this community, the daily tasks most of us perform without a second thought require a level of assistance that isn’t always available, isn’t always timely, and isn’t always dignified.
That’s the gap AI can help close in ways that weren’t possible before. Through Be My AI and our Customer Accessibility Suite, we use AI to help users interpret images, understand their surroundings, and resolve everyday problems more independently. This isn’t about adding convenience to an experience that was already working. It’s about giving people meaningful access to information and autonomy in a world that has, for far too long, been designed around their exclusion.
That’s the distinction at the center of how we think about this. We believe AI’s footprint must be taken seriously, but that AI used to expand independence and accessibility creates a form of human value that has to factor into the conversation.
My view is that the sustainability conversation needs to develop enough sophistication to recognize that AI that removes exclusion is not the same as AI that tweaks with convenience. Both carry costs. They don’t carry equivalent value.
If AI and sustainability are going to coexist – and I believe they can – then responsible development has to stop being the exception and start being the expectation.
That means building leaner systems where possible. Using infrastructure that improves over time. Being honest about tradeoffs rather than retreating into abstraction or greenwashing. And critically, choosing applications where the social return is not just positive, but clear and demonstrable.
This is especially important for companies working in accessibility. Accessibility should never be a marketing theme layered on top of high-consumption technology. It should be a long-term commitment to reducing barriers in ways that are thoughtful, measurable, and genuinely responsible.
Yes. But not automatically, and not by default.
They can coexist when companies are honest about AI’s environmental cost instead of pretending it doesn’t exist. When we stop treating all AI use as morally interchangeable. When we prioritize applications that create real, lasting human value — not just marginal efficiencies.
AI is not sustainable by virtue of being AI. It becomes more justifiable when it’s deployed with purpose.
For us, that purpose is concrete: helping build a world where blind and low-vision people can access information, support, and everyday independence more easily than before. That doesn’t erase AI’s footprint. It doesn’t make the cost disappear. But it does change what that cost is for and the value it brings.
And honestly? That’s the conversation I think matters most. Not whether AI is good or bad in the abstract, but whether the people building it are honest enough, and purposeful enough, to make it worth defending.
