The right to disconnect: a global map of after-hours work

At 10:47 on a Tuesday night, a backend engineer in Dublin gets a Slack message from a product manager in San Francisco. It is not an emergency. It is a “quick question” about a feature spec, the kind that could have waited until morning and almost certainly should have. The engineer stares at the notification, feels the familiar tug to just answer it, and then does something that would have been unremarkable a decade ago and is quietly radical now: she leaves it unread until 9am. In Ireland, she has a legal right to.

That undramatic moment is where a fast-moving global story is playing out. Over the past several years a striking number of countries have passed, or proposed, a “right to disconnect” — a rule that workers may switch off outside working hours without penalty, and that being unreachable after the workday cannot be held against you. France did it first in a meaningful way in 2017. Australia finished phasing one in across 2024 and 2025. The United States, where most of the world’s largest tech companies are headquartered, has nothing at the federal level. For teams that span time zones by design, this is no longer an abstract policy debate. It is the difference in what that 10:47pm ping legally means depending on where it lands.

The thing the laws are reacting to

Before the law, there is the culture, and the culture is genuinely out of hand.

Surveys of knowledge workers keep finding that overtime is the baseline, not the exception. More than four in five desk workers — roughly 84 percent in widely cited 2025 figures — say they work overtime regularly. Only about 36 percent are paid for it. That gap is the whole problem in miniature: the extra hours are real enough to fill everyone’s calendar and inbox, and unreal enough to never reach anyone’s paycheck. The work expanded; the compensation, and the boundary, did not.

It is worse for the kind of work datarekha readers do, because software and data work is structurally always-on. A pull request can be reviewed at any hour, an alert can fire at 3am, and a “can you take a look” travels at the speed of a message, arriving with the soft implication that looking now would be appreciated. Microsoft’s workplace research has described knowledge workers being interrupted roughly every couple of minutes during the day — and the same channels do not politely shut off at 6pm. The boundary between “at work” and “reachable” eroded not because anyone decided it should, but because the default settings of modern collaboration tools quietly erased it.

Companies noticed before governments did, and a few tried blunt fixes. Volkswagen configured its servers, years ago, to stop routing emails to certain employees’ phones outside working hours. Daimler let staff auto-delete inbound mail during holidays, so they returned to an empty inbox rather than a wall of dread. These email-curfew experiments are the corporate ancestors of the laws that followed: an admission that if you leave the boundary to individual willpower, willpower loses.

The legal map, and why it is uneven

What turns this from a wellness anecdote into a regulatory story is that legislatures started codifying it — and they did so at very different speeds and with very different teeth.

France’s 2017 law is the landmark. It did not ban after-hours email outright; instead it required companies above a certain size to negotiate with employees over how the right to disconnect would actually work — a subtler and arguably smarter design than a flat prohibition. From there the idea spread across Europe. Belgium gave public- and then private-sector workers a right to ignore out-of-hours contact. Ireland introduced a code of practice that, while not a hard statute, gives an employee’s refusal to engage after hours real standing in a dispute. Italy wove disconnection rights into its rules on remote and agile work. Each of these is a law or formal code, not a slogan.

Then Australia got the most attention, because it arrived recently and with teeth. Its provisions, phased in across 2024 and extended to smaller employers in 2025, let employees refuse to monitor or respond to unreasonable contact outside working hours, with the national workplace tribunal empowered to step in when disputes escalate. Latin America has its own cluster — Argentina, Peru, and Mexico have each moved on the right to disconnect, often tied explicitly to remote-work regulation. And in 2025 India introduced a Right to Disconnect Bill that grounds the idea in something larger than labor policy: the constitutional right to life, the argument being that a life consumed entirely by work is not one the state should let employers demand.

That framing — disconnection as a dimension of the right to a life — explains why this keeps gaining momentum rather than fading as a fad. It is argued not as a perk but as a boundary on what work is allowed to consume.

A caveat the headlines flatten: these are not the same instrument. Some are binding statutes, some are codes of practice that bite only in a dispute, and India’s is a bill that has been introduced, not a law in force. Lumping them together overstates the uniformity. The honest read is that a real and widening cluster of jurisdictions has made after-hours unavailability a protected act, with meaningful variation in how hard that protection bites.

The American exception, and why it matters to everyone else

The conspicuous gap on that map is the country that exports most of the world’s collaboration software.

The United States has no federal right to disconnect, and state-level attempts have stalled. California’s AB 2751, which would have required employers to set clear expectations about non-working hours, did not make it through. New Jersey floated a similar proposal that went nowhere. So the country whose companies set the de facto norms for distributed tech work — Slack, the always-green status dot, the expectation of fast async response — is also the one least likely to legislate a limit on it.

That asymmetry is not just an American problem, because culture flows along the org chart, not the border. When the headquarters sits in a no-limits jurisdiction and the engineers are scattered across ones with real protections, the headquarters’ habits tend to win by default. The PM in San Francisco who pings at 10:47pm is breaking no law she is subject to; the engineer in Dublin who feels the pressure to answer is the one whose legal right is being quietly eroded by a norm set somewhere it does not apply. The law lives in one place and the expectation travels everywhere.

Why a blanket ban is the wrong instinct

Here is where I want to push back on the easy version of this story, because “just ban after-hours messages” tends to fail in practice, and engineers can see exactly why.

Real systems break at inconvenient times. On-call exists because production does not respect working hours, and a flat ban on after-hours contact collides with the reality that an outage at 2am needs a human now. A rule that has to carve out exceptions for incidents, releases, and time-sensitive work soon resembles the very always-on culture it was meant to replace, now with a layer of guilt about which pings are “allowed.” Boundaries everyone has to violate to do their jobs protect no one; they just move the stress from “I have to answer” to “I have to answer and pretend I didn’t.”

There is a subtler failure too. If the rule is “no messages after 6pm,” the predictable workaround is that people draft at 6pm and feel its weight all evening anyway, or batch a Monday-morning avalanche that is its own kind of dread. The harm was never only the notification sound; it was the expectation of responsiveness the notification carried. You do not fix an expectation by muting a channel — you fix it by changing what people are expected to do when a message arrives.

Designing the boundary instead of decreeing it

So if the blanket ban is theater, what does a humane norm look like for a distributed tech team? The good news is that the better answer is mostly cheaper than the law, and entirely within a team’s control.

Start by separating urgency from time-of-day, explicitly and in writing. What makes 10:47pm corrosive is not the hour; it is the ambiguity about whether it demands a reply. A team that agrees, out loud, that chat and email are asynchronous by default, that the expected response is the next working day in your time zone, and that a genuine page goes through the on-call channel everyone knows means now has done more for its members’ evenings than any auto-delete setting. The boundary becomes about response expectation, not about silencing the tool — exactly the instinct France reached for in requiring negotiation over how disconnection works rather than dictating one rule.

Then make that default real by removing the penalty for using it. The right to disconnect, in every serious version, is fundamentally a protection against retaliation — the promise that not answering after hours will not be remembered against you at review time. A team can grant that without a statute. If the person who silences notifications at night is quietly seen as less committed than the one who replies at midnight, no policy will hold. The lever is making visible after-hours responsiveness a non-signal — neither rewarded nor expected — so that switching off costs nothing.

And treat the time zones as something to design around, not an excuse for round-the-clock presence. A genuinely global team has honest options: a rotating on-call that follows the sun so the 2am burden lands on whoever is awake, a deliberately small overlap window where synchronous work is concentrated, and a firm norm that anything handed off outside that window is genuinely asynchronous. The alternative — every region implicitly on call for every other region’s daylight hours — is how you reinvent the always-on culture the laws were written to stop, with extra steps.

The line you actually get to draw

The map will keep filling in. More countries will pass these laws and the proposals will become statutes, which means anyone on a borderless team will increasingly work alongside colleagues whose right to switch off is stronger or weaker than their own.

But the law was always downstream of the culture, and the culture is the part you can change first. A right-to-disconnect statute is, at bottom, a society formalizing a boundary that good teams could simply choose to honor: that being unreachable after hours is a normal, unpunished, fully professional thing to do. You do not need to wait for your country to legislate the engineer in Dublin’s choice. You can decide, on your own team, that the next-morning reply is the default and the midnight reply is the exception — and then mean it when someone takes you up on it. The most humane version of the right to disconnect is not the one a parliament grants. It is the one a team gives itself, and then refuses to quietly take back.