Why the Free Internet Now Feels Like a Negotiation Every Time You Open a Website

You visit a website for one answer. Before you reach it, the site may ask for permission to track you, your email address, an account, a subscription or another click.
Sometimes the obstacle is a cookie banner. Sometimes refusing tracking requires navigating through several menus. A subscription may take seconds to activate but much longer to cancel. A shopping site may advertise one price and reveal additional charges only at checkout.
Each practice can have a legitimate explanation. Publishers need revenue. Privacy laws require companies to offer users certain choices. Subscriptions are a normal way to pay for journalism and digital services.
But together, these systems have changed the character of the open web. The internet may still be free at the point of entry, but access increasingly comes with a negotiation over personal data, money, time and attention.
The deeper question is whether these demands are necessary to sustain online services, or whether websites have learned to turn inconvenience into a business model.

The Economics Behind the Friction
Brett Caraway, an associate professor at the University of Toronto, links the growing friction to two broader changes: the expansion of privacy rules across different jurisdictions and the concentration of power in the platform economy.
Websites operating across the European Union, Canada, Oceania and other markets must navigate different legal requirements. That can produce more consent notices, privacy settings and compliance processes.
At the same time, companies such as Google, Meta, TikTok and Amazon control large portions of digital advertising, distribution and audience attention. Smaller publishers often depend on these platforms to reach readers, even as the platforms capture much of the advertising value.
Advertising revenue may be growing, but that does not mean every publisher benefits equally. A news organisation that receives much of its traffic through search, social media or video platforms has less control over the relationship with its own readers.
Caraway compares the resulting subscription economy to the shift from one cable bill to several streaming subscriptions. Readers may be willing to pay for quality journalism, but the cumulative effect is a fragmented digital experience in which every publication asks for a separate commitment.
Paywalls, newsletters, memberships and subscriptions can help publishers survive. They can also create a web in which users must repeatedly decide whether a single article, service or feature is worth another payment.
That financial explanation, however, has limits. It may explain why a publisher needs revenue. It does not automatically justify a confusing cancellation process, a preselected option or a checkout screen that hides important costs.
The economics of the internet may explain why websites ask for something. It does not explain every way they ask.

When Bad Design Becomes Manipulation

Dr. Sushila, a professor of law at National Law University Delhi who has contributed to Indian government initiatives on dark patterns, distinguishes ordinary inconvenience from design that interferes with consumer autonomy.
A difficult interface is not automatically a dark pattern. The concern arises when a company intentionally steers users toward a choice they might not otherwise make.
That can involve hiding material information, creating false urgency, preselecting options, making refusal harder than acceptance or adding unnecessary steps to cancel a service.
Indian consumers encounter many of these practices in e-commerce, travel, food delivery and subscription services. They include drip pricing, basket sneaking, false urgency, confirm shaming, forced action, nagging and subscription traps.
A travel site may advertise a low price and reveal additional charges only at the final stage. A shopping platform may quietly add an extra product or service to a basket. A website may claim that only one hotel room remains, creating pressure to buy before the supposed opportunity disappears.
Sushila’s central distinction is that inconvenience alone is not proof of manipulation. The issue is whether a design deliberately interferes with a decision that should belong to the consumer.
As the article puts it, “The question is not whether the interface is irritating. It is whether the design is influencing a choice that should belong to the consumer.”
That is the dividing line between poor design and a dark pattern. A website may be cluttered or badly organised without violating user autonomy. But if it makes acceptance simple and refusal obscure, the problem becomes more than bad usability.

The Cookie Banner Paradox

Cookie banners were introduced to give users greater control over tracking. In practice, they have become one of the most repetitive and confusing parts of browsing.
A 2025 YouGov survey cited in the supplied material found that 82% of Germans were annoyed by cookie banners. Only 14% said they were not annoyed. The same survey found that 36% of respondents did not understand the content of cookie settings, while only 17% consciously selected tracking options.
Alexander Rabe, managing director of the eco Association, summarised the problem this way:
“Cookie banners were originally established to provide greater data protection and control for Internet users. In fact, they mainly result in one thing: consent fatigue.”
The phrase captures a central weakness in the current system. A user may technically have a choice but lack the time, information or patience to exercise it meaningfully.
Cristian Borcea, a professor at the New Jersey Institute of Technology who researches privacy and web tracking, does not argue that cookie consent mechanisms are inherently useless. Users can, in principle, refuse tracking or limit the data collected about them.
The problem is how that choice is presented.
An opt-out option may be hidden inside a settings menu. One button may be larger or more brightly coloured than another. Wording may suggest that the website will not function properly unless tracking is enabled.
Research from Inria privacy researcher Nataliia Bielova and others has highlighted the gap between regulatory expectations and actual user behaviour. Bielova described the disconnect as follows:
“It’s as if there were two separate worlds: the regulatory guidelines on one side, and the user behavior on the other.”
That gap is crucial. A banner may satisfy a formal legal requirement while still being designed in a way that encourages users to accept tracking without understanding the consequences.
A 2026 study discussed in the supplied material found aesthetic manipulation techniques in 38% of cookie banners that were technically GDPR-compliant. That does not mean 38% were illegal. It suggests that legal compliance and meaningful consent are not always the same thing.
Legal scholar Kate Klonick has made an even broader criticism of the model in her Harvard Journal of Law & Technology article, “Ban Cookie Banners”: “Cookie banners do not empower; they weary. They do not inform; they obscure. They do not prevent surveillance; they normalize it.”
The criticism is not simply that banners are annoying. It is that the repeated ritual of clicking through them may normalise surveillance while giving users the impression that they have exercised meaningful control.
Borcea has proposed a more predictable approach: regulators could require a simple, standardised format for cookie banners. That would not necessarily eliminate tracking, but it could reduce the effort required to understand and reject it.

Subscription Traps and the Cost of Saying No

The negotiation does not end with cookies. Subscription services have created another form of digital friction: easy entry and difficult exit.
The Federal Trade Commission’s action against Amazon over its Prime subscription practices brought the issue into sharper focus. The FTC alleged that Amazon used deceptive methods to enrol consumers and made cancellation “exceedingly difficult.” The case involved a reported $2.5 billion settlement in 2025.
The supplied material also cites internal Amazon language describing the company’s practices as “a bit of a shady world.” That phrase, if accurately attributed to the relevant court or regulatory record, illustrates the concern that cancellation barriers can be part of a deliberate commercial strategy rather than an accidental design flaw.
Researchers Amelia Fletcher and her coauthors have described the behavioural economics behind subscription traps. Companies rely on “set-and-forget” defaults and burdensome cancellation procedures, knowing that some consumers will underestimate future costs or fail to cancel in time.
The result is a business model that can earn money from inertia. A customer may no longer use a service but continue paying because cancellation requires another search, another login, a phone call or several additional confirmation screens.
Carter McCants of William & Mary Law School identifies tactics such as hidden cancellation pages and “click to subscribe, call to cancel” arrangements.
The imbalance is obvious. If a company can make signing up almost effortless, it is difficult to justify making cancellation substantially harder. The process may be legal in some circumstances, but it weakens the idea that consumers are making an informed and continuing choice.
The practical principle is simple: cancelling should be at least as easy as subscribing.

The Surveillance Economy

Beneath cookie banners and subscription screens lies a larger economic system built around personal data.
A widely circulated phrase captures the concern: “If a service is free, you are not the customer. You are the product.”
The line is rhetorically powerful, though it is too simple to describe every free online service. Some websites are supported by advertising, some by subscriptions, some by donations and some by a mixture of revenue sources.
Still, the phrase points to a real shift in the relationship between users and platforms. Every search, scroll and click can provide information about interests, habits and likely behaviour. That information may be used to target advertising, personalise content or predict what a user might do next.
Privacy expert Bruce Schneier has argued that the scale of online tracking has continued to expand. Reflecting on the period since his book Data and Goliath, he said: “In the main, nothing has changed since 2015… companies ranging from the large tech monopolies to invisible data brokers are spying on us even more extensively.”
Schneier also challenges the idea that privacy can be protected through consumer choice alone:
“While people realize it, they also realize that they can’t realistically opt out. This is why the notion of consumer choice doesn’t make sense here, and we need a comprehensive privacy law.”
That argument goes beyond cookie banners. It suggests that asking individuals to make dozens of privacy decisions every day may be an inadequate response to an industry-wide system of data collection.
Users may technically be able to refuse tracking, but doing so can require time, technical knowledge and repeated effort. In some cases, refusing may also mean losing access to a service.
Schneier’s long-term view is more optimistic: “In 50 years, I think we’ll view these business practices like we view sweatshops today: as evidence of our less ethical past selves.”
Whether that prediction proves correct, it raises a fundamental question about the present: will future generations see today’s data economy as a necessary stage of technological development or as a period when companies extracted too much from users with too little accountability?

Can Advertising Work With Less Personal Data?

The current internet economy rests partly on the belief that publishers will lose money if personal data is unavailable and targeted advertising cannot be served.
Borcea says existing research supports the concern that restrictions on tracking can affect advertising revenue. But there is an important uncertainty: it is difficult to know what would happen in a realistic environment where targeted advertising and extensive web tracking were broadly restricted.
Advertisers would still have money to spend. Publishers would still have audiences. The market could adapt in ways that are difficult to predict from the current system.
That uncertainty matters because the industry often presents tracking as though it were the only workable foundation for free online content.
Borcea’s research team is exploring privacy-preserving artificial intelligence techniques that could support targeted advertising while limiting the exposure of personal information. He does not present the approach as a ready-made replacement. It would require cooperation among publishers, advertisers, platforms and technology companies.
Another possibility is to move some privacy decisions from individual websites to the browser.
A universal browser-level opt-out could reduce the number of banners and limit consent fatigue. But it would also create a trade-off. A user who does not particularly care about privacy could opt out once and unintentionally prevent publishers from earning advertising revenue from that person, even if the user would otherwise have accepted targeted advertising.
That raises a difficult question about control. Should privacy decisions be made separately on every website, once at browser level or through a combination of both?

The Path Forward

The research points toward several possible reforms. First, cookie banners could be standardised so users do not have to learn a different consent system on every website. Acceptance and refusal should be equally visible, understandable and easy to complete.
Second, regulators could strengthen enforcement against dark patterns, particularly in subscriptions, travel, e-commerce and food delivery. The focus should be on the effect of design on consumer choice, not merely on whether a company includes the required legal wording.
Third, cancellation should be no harder than registration. Companies should not be allowed to use hidden pages, unnecessary calls or repeated retention offers to keep consumers trapped in unwanted subscriptions.
Fourth, privacy-preserving advertising could receive greater attention. The aim would be to help publishers earn revenue without requiring the same level of personal tracking.
Finally, policymakers may need to move beyond the idea that privacy is solely an individual responsibility. Users cannot realistically negotiate every data decision on every website, every day.
The three experts in the article examine different parts of the same system. Caraway explains the financial pressure on publishers as advertising and audience discovery become concentrated around major platforms. Sushila identifies the point at which ordinary friction becomes interference with consumer choice. Borcea explores whether advertising can continue to support the web without requiring the same level of tracking.

None of these solutions resolves the entire problem.

Standardised cookie banners could make consent easier to understand, but they would not solve the financial weakness of smaller publishers. A universal opt-out could reduce the number of consent requests, but it might also reduce advertising revenue. Privacy-preserving advertising could limit the use of personal data, but it would require industry adoption and commercial incentives.
The deeper problem is that the free internet has become dependent on systems that users rarely control: platform distribution, digital advertising, data collection, subscription pricing and interface design.
Publishers need money to produce journalism and maintain websites. Platforms want to preserve their advertising advantages. Regulators want consent to be meaningful. Users want privacy and straightforward access to information.
Those goals are not necessarily incompatible. But under the current model, users often absorb the cost of the conflict. They spend time rejecting cookies, comparing subscription terms, searching for cancellation links and checking whether the price shown at the beginning is the price they will actually pay.
The internet has not stopped being free. It has become conditional.
The bill is often paid in attention, data and cognitive effort. And the most important question for the next phase of the web is whether users are being offered a real choice, or merely being given several ways to agree.