How Do Laws Make Technology Platforms Pay News Publishers?

Laws can make large technology platforms compensate news publishers by requiring bargaining, creating a neighboring right for online use, imposing a levy, or making arbitration the fallback when negotiations fail. These systems do not calculate one universally accepted “price of a link.” They create rules that alter bargaining power, define eligible journalism, and decide what happens if a platform refuses to make agreements.

Why do governments intervene?

News production and online distribution play different roles. Publishers pay reporters, editors, photographers, legal teams, and technical staff to produce journalism. Search engines, social networks, aggregators, and other platforms help audiences discover it, while also controlling important routes to attention and advertising.

Both sides can create value for the other. A link may send a publisher traffic, while headlines and excerpts may keep a platform useful and engaging. The policy dispute begins when bargaining power is highly unequal. A large platform can negotiate across a whole market or change an algorithm; a small publisher may depend on that platform for a significant share of discovery and cannot credibly walk away.

A news bargaining code attempts to change that negotiating position. It does not by itself prove that every display of news causes measurable loss, nor that all publisher revenue problems come from platforms.

Four common policy models

Countries may combine these tools. The legal label matters less than the incentive structure: what behavior triggers liability, which agreements reduce it, and whether refusing to carry news allows a company to exit the system.

Step 1: define which platforms are covered

A law rarely applies to every website or app. Coverage may depend on local revenue, global size, market power, user reach, control over news distribution, or the type of service. Thresholds protect small firms from disproportionate compliance costs, but they can also create cliffs and encourage companies to restructure products around definitions.

Search, social media, professional networks, news aggregators, app stores, and AI assistants do not use journalism in the same way. A rule drafted around links and snippets may fail to address generated summaries, training uses, or answers that satisfy a user without a click.

Step 2: define an eligible publisher

Governments must distinguish journalism from any site that republishes material or claims to be news. Eligibility can involve editorial standards, original reporting, a minimum number of journalists, local operation, revenue limits, correction policies, or registration with a regulator.

Every filter has tradeoffs. A high threshold can exclude startups, freelancers, nonprofit newsrooms, minority-language outlets, and hyperlocal publishers. A loose threshold can attract low-quality sites created mainly to capture subsidies. Good design also prevents the largest media groups from absorbing nearly all available money.

Step 3: specify the bargaining obligation

The law may require good-faith negotiation, disclosure of relevant information, advance notice of major algorithm changes, nondiscrimination, or offers covering particular uses. Private agreements preserve commercial flexibility, but confidentiality makes public evaluation difficult. Two similar publishers may receive very different terms without readers or smaller competitors knowing why.

Policymakers must also decide whether payment reflects content use, the platform's market power, the cost of journalism, the benefit each party supplies, or broader public policy. These are different concepts and they do not produce the same number.

Step 4: create a credible fallback

A duty to negotiate has little force if either side can stall indefinitely. Arbitration supplies a deadline and an outcome. In “final-offer” arbitration, each party proposes a complete offer and the arbitrator selects the more reasonable one. The risk of losing the whole decision is intended to discourage extreme demands.

Arbitration still needs criteria and reliable data. It can favor parties able to hire experts and sustain long proceedings. If the law lacks transparency or appeal safeguards, settlements may reflect litigation capacity as much as journalistic value.

Step 5: stop avoidance

The most difficult design problem is the exit option. If liability applies only when a platform displays news, it may remove links, previews, or entire publisher pages. That response can hurt publisher traffic and make verified information harder to find, even if it avoids the legal payment.

Governments respond in several ways:

• apply obligations based on market position rather than each displayed link;
• use a levy that remains payable even if news visibility falls;
• prohibit retaliation or discriminatory ranking changes;
• allow private agreements to offset a statutory contribution;
• require a minimum number or diversity of eligible deals.

Each countermeasure raises enforcement questions. A regulator must distinguish genuine product changes from strategic demotion and must measure revenue or exposure that a multinational platform may allocate across services.

Australia's incentive model as an example

Australia's 2026 News Bargaining Incentive proposal illustrates a hybrid approach. The finalized bill described in NewTqnia's report would apply a charge equal to 2.5% of qualifying platforms' Australian advertising revenue unless they make enough eligible publisher agreements. The government designed the charge to remain relevant even if a platform reduces news, while giving qualifying commercial deals credit against the obligation.

The proposal also gives additional weight to smaller, regional, and underrepresented outlets. It had not become law at the time of the source article, and AI services were outside its scope. This distinction matters: a finalized government bill is not an enacted and tested regime.

Who actually receives the money?

Payment to a media company is not automatically investment in reporting. A law may therefore tie eligibility or distribution to newsroom employment, original public-interest journalism, local coverage, or audited spending. It may create grants for very small publishers that cannot bargain individually.

Distribution formulas can use audience, journalist headcount, reporting expenditure, geographic service, or equal base grants. Audience-heavy formulas reward reach but favor incumbents. Cost-based formulas may reward inefficient organizations. Equal payments can ignore the scale of actual journalism. A mixed formula is common because no single measure captures public value.

What can go wrong?

• News withdrawal: platforms block or reduce news rather than accept the obligation.
• Concentration: large publishers secure most money and strengthen their position over smaller rivals.
• Opaque deals: confidential contracts hide prices, conditions, and unequal treatment.
• Dependency: newsrooms become reliant on payments a platform or future government can change.
• Editorial influence: commercial relationships may create perceived or real pressure on coverage.
• Innovation barriers: broad rules may burden new aggregators or services that lack dominant market power.
• Traffic loss: fewer links can reduce subscriptions, advertising, and public access to reliable sources.

Are these laws a “link tax”?

The phrase is politically powerful but often imprecise. Some systems attach rights or payment to reproducing headlines and excerpts. Others regulate bargaining with designated platforms or impose a revenue-based contribution whether or not one specific link generated income. Calling every model a link tax can hide these differences.

Ordinary hyperlinks, copied excerpts, full-text reuse, search snippets, social recommendations, and AI-generated answers are technically and legally distinct. A sound analysis asks exactly which act triggers the rule.

What changes with generative AI?

Generative systems complicate the exchange. A search or social link may send a reader to the publisher. An AI answer may synthesize reporting and satisfy the question without a visit. Relevant questions include whether content was licensed for training or retrieval, whether the answer reproduces protected expression, whether sources are attributed, and whether attribution produces meaningful traffic.

Existing bargaining codes may not cover these uses because their definitions predate AI answers or deliberately leave copyright disputes to another process. Extending them is not a simple copy-and-paste exercise: training, retrieval, quotation, summarization, and linking require different evidence and remedies.

How should success be measured?

The mental model

Think of these laws as rules for a negotiation at a gateway, not a meter attached to every article. The platform controls an important route to audiences; publishers supply journalism; the law changes what happens if their bargaining power is unequal. Its success depends not merely on moving money, but on whether the new incentives produce more independent reporting without shrinking public access or locking journalism into a new dependency.