Sustainability and Environmental Impact of Museum Audio Guides
Frequently Asked Questions
Are AI audio guides bad for the environment?
The energy and water used by AI systems gets a lot of attention, but the numbers in context are small. A single hamburger uses more water than 600 years of daily AI queries. BYOD audio guides that run on visitors' existing phones add essentially zero hardware footprint — the environmental cost is the compute, which is a rounding error compared to manufacturing and shipping dedicated devices.
How do BYOD audio guides reduce environmental impact compared to hardware devices?
BYOD guides run on phones visitors already own. There are no dedicated devices to manufacture, ship, charge daily, repair, and eventually dispose of as e-waste. No charging stations drawing power overnight. No fleet replacements every four to six years. The entire hardware lifecycle is eliminated.
What is the e-waste impact of traditional audio guide devices?
A typical museum fleet of 50-100 proprietary devices has a lifespan of four to six years before replacement. That's 50-100 units of consumer electronics — batteries, circuit boards, screens, plastic casings — going to landfill or recycling on a regular cycle. Multiply across the thousands of museums using hardware guides globally, and it adds up.
Do museums care about sustainability when choosing audio guides?
Yes. Sustainability comes up in procurement conversations, especially at publicly funded institutions with environmental reporting requirements. It's rarely the deciding factor, but it's a real consideration — and when the more sustainable option is also cheaper and easier to maintain, it becomes part of the case.
Museums are in the preservation business. That extends beyond collections. Public institutions face real pressure on environmental impact — from trustees, funders, visitors, and increasingly from procurement requirements that ask vendors to account for their carbon footprint.
Audio guides don't usually feature in sustainability audits. They should. The gap between the most and least sustainable approaches is wider than most people assume.
The hardware problem
A traditional audio guide system is consumer electronics, manufactured and distributed at scale within a single building.
Take a mid-sized museum with a fleet of 80 devices. Each unit contains a lithium battery, a circuit board, a screen or speaker assembly, and a plastic casing. Manufacturing those 80 devices means raw material extraction, component fabrication across multiple countries, assembly, packaging, and shipping. The supply chain looks like any other electronics product — because that's what it is.
Once they arrive, the devices need charging. Every night, staff plug 50-100 units into charging stations that draw power for hours. Over a year, that's a real electricity line item. Not enormous, but constant. And it's pure overhead — energy spent not on the visitor experience but on keeping devices ready for the next day.
Then there's the replacement cycle. Audio guide hardware lasts four to six years before the batteries degrade, screens fail, or the technology falls behind. At that point, the entire fleet gets replaced. The old devices become e-waste. Some get recycled properly. Many don't.
Between replacements, there's attrition. Broken units, lost units, units with failing batteries. A 5-15% annual replacement rate is standard. That's a steady stream of small electronics going into the waste stream year after year.
None of this is malicious. It's just the physics of maintaining a fleet of dedicated electronic devices. The environmental cost is baked into the model.
Printed guides aren't better
The obvious low-tech alternative — printed guides, pamphlets, laminated sheets — has its own problems.
Paper guides get thrown away. Even laminated versions degrade and need reprinting. A museum that updates its exhibitions regularly is reprinting guides multiple times a year. Multilingual support means multiplying the print run by every language offered. A temporary exhibition with a six-month run might go through thousands of printed guides that end up in bins on the way out.
And unlike a digital system, there's no way to update a printed guide without reprinting it. A factual correction, a changed gallery layout, a new temporary exhibition — each requires a new print run and disposal of the old stock.
The per-unit environmental cost is small. A single pamphlet is nothing. But aggregated across millions of museum visitors globally, the paper waste is substantial.
BYOD changes the equation
Bring-your-own-device audio guides work on visitors' phones. The phone is already manufactured, already shipped, already in the visitor's pocket. The audio guide adds zero hardware to the world.
No devices to produce. No batteries to charge overnight. No fleet to replace every few years. No e-waste at end of life. No charging stations. No packaging, no shipping containers, no warehousing.
The environmental footprint of a BYOD audio guide is, practically speaking, the energy used to run the software — both on the visitor's phone (negligible, less than watching a YouTube video) and on the servers delivering the content.
That's not a marginal improvement. The entire hardware lifecycle — manufacture, ship, charge, maintain, replace, dispose — just doesn't exist in a BYOD model.
The AI energy question
This is where the public conversation has gone somewhat sideways.
AI-powered audio guides use large language models. Running those models requires compute. Compute requires electricity. Electricity requires power plants. Data centers use water for cooling. Headlines have seized on this chain to argue that AI is an environmental disaster.
The absolute numbers sound alarming in isolation. Data centers use millions of gallons of water. AI workloads consume megawatts of power. If you stop there, it looks bad.
But the numbers only mean something in comparison. And the comparisons tell a different story.
Semi Analysis published a detailed breakdown comparing AI data center water usage to everyday activities. The finding that sticks: the largest AI data center in the world uses roughly as much water as two and a half In-N-Out Burger locations. Not the entire chain. Two and a half restaurants.
Put differently: the water footprint of producing a single hamburger — about 245 gallons when you account for the full supply chain of raising cattle, growing feed, and processing — exceeds the water cost of using an AI assistant 30 times a day, every day, for over 600 years.
Six hundred years. For one burger.
The energy picture is similar. A single AI query uses a fraction of a watt-hour. Streaming a Netflix show for an hour uses orders of magnitude more energy than asking an AI guide a dozen questions during a museum visit. The visitor's phone screen being on for the duration of the tour likely draws more power than the AI inference serving their questions.
None of which means AI energy use is zero, or that it doesn't matter at scale. But for any individual museum evaluating whether an AI-powered audio guide is environmentally responsible, the answer is clear. The compute footprint of serving AI-generated narration to visitors is a rounding error next to manufacturing, charging, and disposing of a fleet of dedicated hardware devices.
What Musa's footprint actually looks like
Musa is a BYOD AI audio guide system. Visitors use their own phones. There are no dedicated devices, no charging infrastructure, no hardware lifecycle.
The environmental cost of running a Musa guide at your museum is the server-side compute for AI generation — the electricity and cooling required when a visitor asks a question or the guide generates narration. That cost is real, but as the numbers above show, it's tiny in absolute terms. A busy museum generating thousands of AI interactions per day uses less energy on those interactions than the museum's lighting system uses in a few minutes.
In practice, Musa's environmental footprint is essentially identical to any other BYOD web application your visitors might use. It's in the same category as checking a museum's website on your phone, looking up opening hours, or reading a Wikipedia article about an exhibit. Normal phone usage. Nothing additional manufactured, nothing additional disposed of.
The circular economy angle
There's a broader framing. Traditional audio guides follow a linear model: extract materials, manufacture devices, use them for a few years, dispose of them, start over. Every replacement cycle restarts the chain.
BYOD audio guides sit naturally within a circular economy framework. They add no new physical products to the cycle. The phone the visitor uses was going to be manufactured regardless. The guide is software — it doesn't consume physical resources beyond the marginal electricity for compute and data transfer. When the guide needs updating, it's a software change. No old products to dispose of, no new ones to ship.
For museums that report on sustainability metrics, switching from hardware audio guides to a BYOD solution is a concrete, quantifiable reduction in environmental impact. It shows up in annual reports — the audio guide fleet was never a huge emitter, but eliminating it entirely is clean and measurable.
Why this matters for procurement
Sustainability isn't usually the reason a museum switches audio guide systems. Cost, visitor experience, and content flexibility drive most decisions. But sustainability is increasingly part of the conversation.
We've had it come up in sales. Museums — especially publicly funded ones — ask about environmental impact during procurement. Some have formal sustainability criteria in their vendor evaluation. Others just want to know they're not adding unnecessary waste.
The good news: the most sustainable option is also the most practical one. BYOD guides are cheaper than hardware (no devices to buy), easier to maintain (no fleet to manage), more flexible (updates are instant), and produce less waste (none, really). Sustainability and functionality point in the same direction here.
Being honest about tradeoffs
A few honest caveats.
BYOD guides depend on visitors having smartphones. Most do — smartphone penetration among museum visitors is above 95% in most markets — but not all. Some institutions keep a small number of loaner devices for visitors without phones, which reintroduces a modest hardware element. It's a fraction of a full fleet, but it exists.
AI compute isn't carbon-free. The electricity powering the data centers comes from whatever grid mix exists in the region. As grids shift toward renewables, the footprint drops automatically, but today it's not zero. It's just very, very small per interaction.
And manufacturing smartphones has a massive environmental cost. But that cost exists whether or not the phone is used as an audio guide. The guide doesn't cause the phone to be manufactured. It takes advantage of hardware that's already in circulation.
What to take from this
If you're evaluating audio guide systems and sustainability is on the checklist, the math is straightforward. Hardware devices carry a real environmental cost across their entire lifecycle. BYOD eliminates that cost almost entirely. AI compute adds a marginal footprint that is, by any reasonable comparison, negligible.
The scariest-sounding numbers about AI energy and water usage dissolve when you put them next to everyday activities. A museum serving AI-generated audio content to thousands of visitors uses less water than the staff cafeteria. Less energy than the gift shop's lighting. The panic about AI's environmental impact is real, but the scale is wrong — at least for applications like this.
If sustainability matters to your institution, the audio guide decision is one of the easier ones. The option that's better for visitors is also better for the environment. That doesn't always happen. Take the win.