Recipes for
a healthier
digital

It is Thursday afternoon. You open your inbox: 3,847 unread messages. At least three thousand you will never open. But they are all there, on a server somewhere in the world, consuming electricity twenty-four hours a day: even while you sleep, even while you are on holiday, even when you have not opened Gmail for a week.

It is not a problem of untidiness. Every message takes up physical space on a server that needs electricity, cooling, maintenance. An infrastructure constantly on to keep purchase receipts from 2019, notifications from apps you no longer use, and newsletters you have never read.

Multiplied by the 333 billion e-mails sent every day around the world, the problem takes on a scale that is hard to ignore. Our inboxes are tiny personal data centres: and data centres are among the largest energy consumers on the planet.

There is a moment of pure pleasure in peeling the cellophane off a brand-new smartphone. The object is cold, precise, immaculate: it looks born from nothing. In reality it comes from very far away.

Every smartphone is the result of a very long chain: mines in the Congo, smelters in China, factories in Vietnam, containers crossing oceans. 80% of the total environmental footprint has already happened before the courier rings your doorbell.

The answer is not to give up the smartphone. It is to keep it longer. Every extra year of use dilutes that mountain of already-emitted emissions, and pushes back the environmental cost of the next purchase.

On the sofa, favourite series, screen in the dark. It feels like the most still moment of the day. But at that very instant, thousands of kilometres of cables, dozens of routers and entire buildings full of servers are at work, just for you.

Every second of video travels an invisible path: from the data centre to the fibre optic, to the internet provider, to the home WiFi antenna, all the way to the pixel on the screen. Every step consumes electricity. Far more than one thinks.

The problem is scale: Netflix alone serves over 200 million subscribers around the world every day. Video streaming on its own accounts for 60% of all global internet traffic: and that share grows every year.

When we click “Accept all” we do it almost by reflex: less than a second, then we move on. But that choice has consequences that last years.

In that moment we give the green light to dozens of companies (often more than a hundred per single site) to gather information about us. This is aggregated, resold, used to build an ever more precise profile: where you have been, what you have searched for, who you talk to, how much you are willing to pay.

It is not just about targeted advertising. It is about knowing our vulnerabilities in advance, and using that knowledge to steer our decisions, often without our noticing.

Open your computer. How many applications have you installed that you have not used for months? Probably more than you think. Each one runs silently in the background: updating, syncing, sending data to the manufacturer’s servers. All of this consumes real electricity, even when you are doing nothing visible.

A Microsoft study found that 77% of installed business software is never used. Not “rarely used”: never. And yet it keeps running, updating, consuming CPU cycles and bandwidth like an engine left running in the garage.

Choosing lightweight, open-source software is not an ideological choice: it is a concrete choice with measurable impact. Free software tends to be more efficient, longer-lived, and has no commercial incentive to push needless updates that bloat devices and accelerate obsolescence.

Try to remember the last time you waited for something, even just a few minutes, without looking at your phone. A red light, the lift, the three minutes while the coffee cools. For most people the answer is “I don’t remember”. Not because moments of waiting no longer exist: they do, exactly as before. It is that the hand goes to the pocket before anything one could call a conscious decision has even formed.

It is not about laziness or a lack of willpower: it is the predictable outcome of a system designed with great precision to produce exactly this behaviour. The smartphone is the most sophisticated device ever built to capture and hold human attention; the applications it contains are the product of billions of euros of research aimed at a single goal: making sure you keep using them for as long and as often as possible.

Italians check their smartphone on average 160 times a day: many of these unlocks last less than thirty seconds and end without anything meaningful having been done. Every notification that breaks the flow of work takes the brain between fifteen and twenty minutes to return to its previous level of concentration; according to the available research, up to 40% of the productive day of a connected worker is eaten away by digital interruptions.

Open any news site. While the page loads, dozens of invisible things are happening: tracking scripts activating, uncompressed images arriving from distant servers, unused plug-ins consuming CPU. That site weighs perhaps 5 megabytes. Every time you go there. A homepage full of trackers is the digital equivalent of leaving every light in the house on when you go out.

The problem does not only concern users: it concerns those who design sites, those who choose hosting, those who install fifteen plug-ins when three would do. But as users we too can make choices that reward the most sober services, and flag for others when a site is needlessly heavy.

Tools like websitecarbon.com let you measure the footprint of any site within seconds. Sites hosted on servers powered by renewables emit up to 90% less. The Green Web Foundation maintains a public, verified registry of providers with green hosting.

Imagine a company that uses dozens of digital tools every day (e-mail, videoconferencing, management software, remote storage) but has never measured how much all of this weighs. It does not know how much energy its IT consumes, has no clear policies on customer data, does not know the terms of the contract with its remote-services provider. This is the norm, not the exception.

Every organisation that uses digital tools produces an impact: on how much data it collects, how it protects it, how much energy the infrastructure consumes, whether employees are exposed to techno-stress, whether the artificial intelligence systems in use amplify discriminatory bias. Corporate Digital Responsibility (CDR) is the set of values and practices to manage all of this consciously.

Sloweb argues that CDR should be integrated into the business model, not added as decoration. An organisation that handles customer data responsibly, chooses remote-storage providers running on renewable energy, and trains staff on the conscious use of the digital obtains concrete benefits: more trust, lower legal risks, real savings.

You type a question. In less than a second an articulate, fluent, seemingly reasoned answer comes back. It looks like magic. In reality it is the result of a process that has consumed enormous resources: training GPT-3 emitted the equivalent of 550 tonnes of CO₂. Every single query to a large model consumes 10 to 100 times more energy than an ordinary web search.

But the cost is not only energy. The data centres of artificial intelligence consume water in huge quantities for cooling. And behind every model there is an invisible army of workers (often in developing countries) who label data, moderate violent content, correct errors: precarious, badly paid, psychologically wearing labour.

This does not mean artificial intelligence should be avoided. It means it should be used with awareness: choosing smaller, more efficient models when they are enough, preferring tools that are transparent about their impacts, not delegating to artificial intelligence tasks that a human can do better and with fewer resources.

Your mother dies. Her iPhone is locked. Inside it there are years of photos, voice messages, a digital diary you did not even know existed. You try calling Apple: they ask you for a death certificate, a notarised document, proof that you are the rightful heir. Months of waiting, forms, refusals. And in the end, in many cases: nothing.

It is not a rare case. Every year around 600,000 people die in Italy. Almost all of them with a smartphone, one or more e-mail accounts, years of photos in remote archives, active digital subscriptions. And yet less than 5% have ever set up a digital-legacy tool: and Italian law does not yet offer adequate safeguards.

There is a paradox few are aware of: those films on iTunes, those e-books on Kindle, that playlist curated over twenty years: they do not legally exist. When you buy digital content, you do not purchase the object: you purchase a personal licence that expires with you. Your heirs will inherit your physical devices, but not the legal right to the content you paid for.

The paradox of the digital is that it seems immaterial. Files weigh nothing, e-mails do not smoke, videos do not burn fuel. The digital sector as a whole today produces around 4% of global CO₂ emissions, more than world civil aviation. And, unlike aviation, it is growing exponentially.

The environmental footprint is spread across three levels: hardware production (the most impactful phase, with the extraction of rare minerals and the electronic manufacturing industry), the energy consumption of data centres (around 30% of digital emissions, and rapidly growing because of artificial intelligence), and the data-transmission network.

Solutions exist, and some are already available today. Green hosting, frugal software design, right-to-repair policies, energy-efficiency standards: not just individual acts are needed, but infrastructures, regulations and a professional culture that turns the digital from a problem into part of the climate solution.

Somewhere in Kenya, a young man looks at images all day long. Images of violence, of abuse, of content no one should see. He classifies them, labels them, approves or rejects them. For a dollar or two an hour. This labour, invisible, unrecognised, often traumatising, is the foundation on which the artificial intelligence systems we use every day operate.

But the social impact of the digital does not stop there. It concerns algorithmic hooking: platforms designed to maximise attention time using the same psychological techniques as slot machines. The biases of algorithms that discriminate in hiring, in bank loans, in criminal sentences. The digital divide that excludes the elderly, the poor, people with disabilities.

Sloweb proposes a digital literacy that goes beyond knowing how to “use” a tool: understanding the mechanisms of persuasion, recognising patterns of manipulation, consciously choosing how much time one gives up to the algorithms. A choice, not a compulsive use.

Amazon, Microsoft, Google, Alibaba, Salesforce. Five names control almost the entirety of the remote infrastructure on which the digital services of governments, hospitals, universities and businesses worldwide run. This concentration is not the result of free competition: it is the product of an economic model that rewards platform monopoly, where whoever arrives first builds lock-in networks that are hard to dismantle.

The impact on the economy is profound. Artificial intelligence is accelerating the replacement of routine jobs at unprecedented speed, in sectors that affect millions of people. Digital disinformation destabilises democratic processes worldwide. The concentration of data creates unprecedented power asymmetries between those who own the infrastructure and those who live inside it.

Solutions require collective political action. But individual choices, too (which remote infrastructure to use, which platform to depend on, what software to adopt at work), are concrete political acts that contribute to building or dismantling these monopolies.