In August 2023, Meta stopped distributing news to Canadian users of Facebook and Instagram. It was responding to the Online News Act, which required large platforms to compensate publishers whose content circulated on their services. No Canadian lost the legal right to say anything. Individuals could still post opinions, rumours, and commentary freely. What disappeared was a distribution channel for organisations that produce reporting, and with it a measurable share of their audience.

The episode exposes a distinction that public debate routinely collapses. Free speech and free press are often treated as one principle, and constitutional texts sometimes place them in the same clause. But they describe different things. Speech is a capacity individuals possess. The press is an activity that institutions sustain. That activity requires capital, labour, legal defence, and reliable access to audiences. Harold Innis argued that every communication medium carries a bias, favouring certain kinds of knowledge and certain holders of power. The legal and economic frameworks built around media carry biases of their own.

Treating speech and press as a single problem produces policy that protects the wrong thing. A society can maximise individual expression while its capacity for verified, sustained reporting erodes. To see why, we have to look past rights frameworks to the economic and infrastructural conditions that make journalism possible at all. Three dimensions matter most: what production costs, what independence requires, and how platforms reshape both.

Individual speech has a low marginal cost. A person with an opinion and access to a public square, physical or digital, can exercise the right with almost no resources. Legal protection is the main thing required. Remove the prohibition, and speech largely takes care of itself.

Journalism has a different cost structure. An investigation into municipal procurement fraud might require months of a reporter's salary, public records requests, data analysis, editorial review, and legal vetting before anything is published. Most of these costs are fixed and front-loaded. They must be paid whether the resulting story attracts a large audience or a small one.

This is why news has historically depended on cross-subsidy. Classified advertising, display ads, and bundled content such as sports, weather, and listings generated revenue that funded expensive, low-margin reporting on courts and councils. The bundle was an economic architecture, not just an editorial choice. Digital platforms then unbundled these services, moving classifieds to specialised marketplaces and advertising to targeted networks. The subsidy that sustained accountability reporting lost its source.

The consequence is visible in what researchers call news deserts: regions where no outlet systematically covers local government. Residents there keep their full speech rights. They can attend meetings, post criticism, and organise. Yet research has linked declining local coverage to higher municipal borrowing costs and lower civic participation. The right persists, but the institutional function does not.

Press freedom, then, is partly a question of whether a viable business model exists. A legal guarantee against censorship means little when no organisation can afford to produce the reporting that censorship would target. Silence produced by market failure and silence produced by state prohibition look different in law. In effect, they can be very similar.

Takeaway

A right that nobody can afford to exercise is formally intact but practically absent. Press freedom has to be measured by the capacity to produce journalism, not only by the absence of restrictions on it.

Investigating powerful actors requires more than resources. It requires that those resources not come with conditions set by the parties under scrutiny. Independence is a structural property. It is determined by ownership, revenue sources, and legal protections, not by the integrity of individual journalists.

Every funding model carries its own dependencies. Advertising-supported outlets face pressure, rarely explicit, to avoid alienating major advertisers. State-funded broadcasters depend on budgets controlled by the governments they cover. Billionaire ownership can stabilise finances while concentrating editorial leverage in one person whose business interests extend far beyond media. Subscription models reduce some dependencies, but they tend to orient coverage toward audiences able and willing to pay.

Durable independence usually rests on insulation mechanisms rather than on finding a pure revenue source. Public broadcasters with multi-year charter funding, arm's-length boards, and statutory editorial protections are harder to pressure than those funded annually at ministerial discretion. Trust ownership, such as the Scott Trust arrangement behind The Guardian, separates financial control from editorial direction. Nonprofit newsrooms diversify their donors so that no single funder becomes indispensable.

Legal infrastructure matters just as much. Shield laws protecting confidential sources, anti-SLAPP statutes limiting strategic lawsuits, and freedom-of-information regimes all work as institutional supports. These protections are designed for the press as an activity. An individual speaker rarely needs source confidentiality, but an investigative newsroom cannot function without it. And because institutions have assets worth targeting, defamation litigation can be used as a weapon against them.

Seen this way, press freedom resembles a supply chain more than a single right. It depends on revenue streams, governance arrangements, legal shields, and access to information, and each can be weakened independently. A government can leave the constitutional guarantee untouched while degrading the press through how it allocates state advertising, regulatory pressure on owners, or selective access to officials.

Takeaway

Independence is not a character trait of journalists but a design property of institutions. Ask who can withdraw a newsroom's money, access, or legal protection, because that is where its vulnerability lies.

Digital platforms complicate the distinction because they host individual and institutional expression in the same interface. A post from an investigative outlet and a post from an anonymous account appear in identical formats in a feed and pass through the same ranking systems. The cues that once signalled institutional publishing, such as the masthead, the broadcast schedule, and the print edition, are flattened.

This produces a paradox. Individuals gained publishing capacity that once required institutional backing, and a single person can now reach millions. Institutions, meanwhile, lost control over distribution and became dependent on algorithmic systems they neither own nor fully understand. Speech became more like publishing, and publishing became more like speech, and both now answer to the same intermediary.

Platforms also introduce a third actor whose decisions resemble neither state censorship nor traditional editorial judgement. Content moderation and ranking determine visibility at scale. When a platform deprioritises news, as Meta did in Canada and later more broadly, it exercises gatekeeping power without the legal status of a publisher or the accountability of a state. In many jurisdictions, liability protections for intermediaries reinforce this ambiguity.

Tiziana Terranova's analysis of free labour offers a useful frame. Platforms extract value from content contributed without direct payment, whether it comes from individuals or newsrooms. For individuals, the exchange may be acceptable, since posting costs little. For news organisations, it means producing expensive content whose distribution value is partly captured by an intermediary that bore none of the production costs.

Regulatory responses reflect the confusion. Bargaining codes such as those in Australia and Canada treat news as a distinct category that deserves compensation. Content moderation debates usually frame the problem as one of individual speech. Both approaches are partial. Platforms are infrastructure for speech and infrastructure for press at the same time, and the interests of the two are not always aligned.

Takeaway

Platforms did not merge speech and press. They placed both under a shared intermediary whose incentives serve neither by default, and policy that addresses one function while ignoring the other is likely to misfire.

The distinction between free speech and free press is not a legal technicality. It marks the difference between a capacity that law can largely secure on its own and an activity that requires ongoing economic and institutional support.

For those shaping media systems, this points to different diagnostic questions. Speech protections ask whether expression is prohibited. Press health asks whether reporting is being produced, by whom, how it is funded, and through which channels it reaches people. A society can score well on the first and fail the second, and that failure may stay invisible until an important story goes uncovered.

Innis observed that changes in communication infrastructure redistribute power, often before institutions notice. The present moment fits that pattern. Protecting expression remains essential. But sustaining the press is an engineering problem as much as a rights problem, one that involves business models, governance, and distribution architecture, and it deserves deliberate design.