Key Takeaways
Treat personal data as political power, not harmless exhaust.
Data grants leverage. A phone’s location, a search history, or a record of purchases can reveal far more than isolated facts. Combined, these details help institutions anticipate what people may do and shape what they see, buy, or believe. Véliz’s central claim is that privacy is power because controlling access to personal information limits other people’s ability to influence and govern someone’s life.
The bargain is unequal. Google and Facebook may not hand users’ data directly to advertisers, but they can sell access to people’s attention and the ability to target them. The people supplying the information often have little idea who is using it, for what purpose, or how to object. Privacy is therefore not merely concealment. It is a way to resist one-sided control.
What's striking is the shift from treating data as a commodity to seeing it as a source of bargaining power. A company that can predict a person’s choices may influence them without issuing an order; the pressure can feel like personal preference rather than external control. This makes privacy relevant even when no secret is exposed. The argument is strongest where institutions have persistent access and conflicting incentives. It should not imply that every use of data is domination: the key question is who sets the terms, who benefits, and whether the affected person can meaningfully refuse.
Assume anonymous records can identify you when data sets are combined.
Small clues identify people. Removing names from a database does not reliably make it anonymous. Research cited by Véliz found that birth date, gender, and ZIP code could identify 87 percent of Americans. In another study, four location-and-time points uniquely identified 95 percent of people in a large mobility database. A person’s movements can function like a fingerprint.
Cross-checking makes it worse. Researchers re-identified people in a supposedly anonymous Netflix ratings release by comparing movie preferences and timestamps with public information. Similar matching can connect location, purchases, and online activity to a real identity. The practical lesson is to treat data as identifying when it can reasonably be linked to other records, not only when a name is attached.
This challenges a common security instinct: remove the obvious identifier and assume the risk is gone. In networked systems, identity often emerges from combinations, not single fields. That is why privacy assessments need to consider what outside data could be joined to a record, including data that does not exist yet. Still, re-identification risk varies by context, dataset, and access controls. The useful principle is not that every anonymous dataset is immediately identifiable, but that anonymity is a claim requiring evidence and continuing review.
The surveillance economy grew through choices, not technological destiny.
A business model took hold. Google’s search engine began with a system for ranking useful results, but the company needed revenue. It developed advertising that used people’s searches and activity to personalize ads, then expanded its reach across websites. Google’s revenue rose from $19 million in 2000 to $3.2 billion in 2004. Other firms followed the profitable model, making personal-data collection a routine part of the internet.
A political opening followed. In 2000, the US Federal Trade Commission proposed legislation giving people information, choices, access, and security protections for online data. After the September 11 attacks, government priorities shifted sharply toward security, and that effort stalled. Véliz argues that privacy was also wrongly portrayed as outdated, a story that served businesses whose profits depended on collecting more information.
The origin story matters because it replaces inevitability with institutional responsibility. Search, advertising, and data collection could have been arranged differently; policy choices helped determine which business models became dominant. This is a useful corrective to narratives that treat technology as an unstoppable force. Yet the history also shows how incentives compound: once firms build around tracking, competitors and publishers face pressure to follow. Changing course therefore requires more than asking companies to behave better. It requires altering the rewards and constraints built into the market.
Treat stored personal data like hazardous material, not a free asset.
Data creates lasting risk. Véliz compares personal information to asbestos: it can be useful, but it can also cause severe harm when stored, exposed, or repurposed. In the Ashley Madison breach, information about more than 30 million users was published, leading to blackmail, public humiliation, broken relationships, and suicides. A stolen identity can also harm someone with no embarrassing secret at all, as Spanish woman Ramona María Faghiura experienced while repeatedly trying to prove she was not responsible for crimes committed in her name.
History raises the stakes. In Nazi-occupied Europe, detailed Dutch population records helped authorities identify Jewish residents. In France, weaker records and resistance by official René Carmille helped prevent some data from being used for persecution. The comparison shows why collecting less, and being able to delete it, can matter beyond individual privacy.
The hazardous-material analogy makes an important point about delayed harm: data may seem harmless while conditions are stable, yet become dangerous when attackers, institutions, or political regimes change. It also highlights the asymmetry between collecting data, which is easy, and retracting it, which is often impossible. The historical contrast is powerful, though no single factor explains the different outcomes in the Netherlands and France. Its broader lesson is precautionary: systems should not accumulate sensitive records without a compelling need, safeguards, and a credible plan for eventual deletion.
Protect privacy as shared civic infrastructure, not individual property.
One person’s data implicates others. A DNA test can reveal information about parents, siblings, children, and distant relatives who never agreed to participate. Likewise, sharing contacts or location can expose other people. Véliz argues that personal data is not purely personal, so treating it as an asset that one individual may freely sell ignores its effects on others.
Democracy needs shared facts. Cambridge Analytica obtained data connected to about 87 million Facebook users and used psychological profiles to target political messages. Researchers debate how much such campaigns changed election outcomes, but the method could show different, even contradictory, content to different audiences. When citizens cannot see what others were told, public scrutiny and common debate weaken. Privacy protects not only individuals, but also the conditions for collective self-government.
This reframes privacy as a collective-action problem, like clean air: one person’s choice can impose risks on people who did not consent. The analogy explains why individual data ownership alone cannot solve every privacy problem. At the same time, collective protection must leave room for legitimate, consent-based uses, especially in research and healthcare. The hard question is how to distinguish shared risk from paternalism. A strong policy should protect those exposed by someone else’s disclosure while preserving transparent processes for uses that serve a clear public benefit.
Reject mass surveillance sold as a guarantee of safety.
More collection can mean more noise. After September 11, US agencies expanded broad surveillance programs. An FBI review found that only 1.2 percent of tips from one program between 2001 and 2004 were useful under its criteria; a later review found no tips making a significant contribution. Véliz’s point is not that targeted investigation never helps, but that collecting everyone’s data can bury useful leads in irrelevant information while weakening civil liberties.
Emergencies invite permanent exceptions. In a crisis, governments and companies can present surveillance as necessary before its benefits are established. Measures introduced as temporary may persist after the immediate danger fades. Because terrorism and epidemics cannot be reduced to zero risk, policy should demand evidence, limits, oversight, and a clear end point rather than treating safety as an unlimited justification.
The useful distinction is between focused investigation and indiscriminate collection. The latter promises certainty against rare threats, but that promise can encourage a false choice between privacy and safety. The cited FBI figures support skepticism about bulk collection, though they do not establish that every surveillance technique is ineffective. Good governance should compare a measure’s demonstrated benefit with its rights costs and security vulnerabilities. Crisis policy especially needs review dates and public accountability, since fear can make extraordinary powers politically easy to grant and difficult to withdraw.
More personal data does not automatically produce more progress.
Data is not a cure-all. Véliz challenges the claim that privacy protections necessarily block innovation. AlphaZero, an AI system that mastered Go and other games, learned through self-play rather than personal data. In medicine, highly publicized AI projects often failed to demonstrate benefits: a review of about 20,000 medical AI studies found only 14 sufficiently rigorous for clinical testing. These examples do not show that digital tools cannot help, but they do challenge the assumption that larger personal-data stores guarantee better results.
Useful research needs safeguards. Medical information can support important discoveries, but people should know what data is used, consent where possible, and receive protection and fair benefit. During the early coronavirus response, contact-tracing apps relied on imperfect proxies for infection; testing and public-health measures could be more direct. The question is not whether data may ever be used, but whether the method works and the risks are justified.
This is a valuable distinction between data volume and knowledge quality. A large dataset can amplify weak measurements, biased assumptions, or an ill-defined clinical goal. The book’s examples also encourage a more careful standard for technological claims: evaluate outcomes, not demonstrations or promises. However, AlphaZero is a narrow example from game-playing, not a general model for medical AI, and some research does depend on sensitive information. The strongest conclusion is therefore not that less data is always better, but that personal data should not be treated as a substitute for sound methods.
Make data collection exceptional, limited, and accountable by design.
Change the default. Véliz proposes ending personalized advertising and the commercial trade in personal data, including data shared for business advantage. Collection should be opt-in and limited to information genuinely necessary to provide a service, rather than information gathered just in case it proves profitable. Contextual ads, based on the page or search topic rather than a person’s profile, offer one alternative.
Make institutions bear responsibility. Companies entrusted with personal data should owe duties of loyalty and care, similar to professionals who must put clients’ interests first. Strong cybersecurity standards, narrow limits on sensitive inferences, and deletion plans would reduce the risk of data accumulating indefinitely. Differential privacy, which adds controlled statistical noise to protect individuals while preserving useful aggregate results, is one possible technique, though it is not suitable for every task.
The proposals address the incentive problem at several levels: reduce collection, limit commercial circulation, and make custodians responsible when they misuse or lose data. This is more coherent than relying on unread consent notices to repair an unequal relationship. The main implementation challenge is defining necessity and enforcing duties across different sectors without obstructing legitimate services. The book’s emphasis on privacy-preserving methods helps avoid a simple ban-versus-innovation debate. Regulation works best when it sets enforceable boundaries while leaving room for safer ways to achieve a genuinely useful purpose.
Reduce your digital footprint with practical friction, not perfect secrecy.
Start with high-impact choices. Share less, ask before posting other people’s photos, and avoid uploading children’s images without their parents’ permission. Prefer devices without unnecessary microphones or internet connections. Use encrypted messaging, a privacy-conscious search engine, and an ad blocker; avoid reusing passwords, and consider a password manager. Keep work and personal accounts separate, and delete information that no longer needs to be online.
Small refusals add up. Decline unnecessary requests for an email address or data access, and choose privacy-friendly services when practical. Obfuscation means deliberately giving ambiguous or misleading information to disrupt tracking when a company has no legitimate claim to accurate details. It is not appropriate for every setting, such as a justified request from a tax authority. These steps cannot guarantee privacy, but they reduce exposure and signal demand for better options.
The practical advice is strongest when treated as risk reduction rather than personal blame. Individuals cannot inspect every database or secure the infrastructure of every service they use; the burden cannot fairly rest on users alone. Still, friction can interrupt automatic disclosure and create better habits. A useful approach is to prioritize the most sensitive activities and the largest exposures instead of chasing perfect privacy. That keeps protective behavior manageable, while recognizing that tools and settings change and cannot replace enforceable institutional standards.
Turn privacy habits into public pressure, not private retreat.
Individual action has a civic role. Véliz urges readers to contact representatives, support independent journalism, request access to data held by companies, and complain when services use invasive practices. People working in technology can push for privacy safeguards or oppose projects that risk harming users. Choosing better alternatives also helps demonstrate that privacy can be part of a viable service, not merely a cost.
Organize for system change. Privacy depends on cooperation, so people can discuss it with friends, create no-recording spaces for sensitive conversations, and support collective action by workers and citizens. Personal measures are imperfect, but they communicate that surveillance is not automatically acceptable. The larger aim is political: laws, oversight, and safer technology should make privacy the default, so individuals do not have to become security experts just to participate in ordinary life.
The book avoids making privacy a purely technical hobby. Its broader claim is that norms and markets shift when people coordinate their expectations, much as workplace protections required both cultural pressure and legal change. That is a more durable strategy than asking every person to outsmart every tracker. Public pressure also needs specific demands, such as limits on collection, independent oversight, and meaningful remedies, or it can dissolve into vague distrust. The balance matters: personal precautions are useful now, while collective rules determine whether those precautions remain necessary.
Analysis
Privacy Is Power is an argument about political economy as much as personal boundaries. Its narrative moves from the ordinary accumulation of data to the institutions that convert it into predictive and persuasive power, then to the risks created when data spreads beyond its original context. The central mechanism is asymmetry: organizations can infer, classify, and act on people, while individuals often cannot see the inferences, contest decisions, or withdraw information. That asymmetry has consequences for autonomy, equality, intimacy, public debate, and democracy.
The book’s most durable contribution is its insistence that privacy is not simply a preference for secrecy. It is a condition that limits arbitrary influence and preserves room for people to experiment, disagree, and change. The historical examples sharpen the point: records that seem administrative or commercially useful can become instruments of discrimination, coercion, or persecution under different political conditions. The argument also usefully rejects technological inevitability. The advertising model, default data collection, and extensive retention are policy and business choices, not unavoidable properties of digital tools.
The proposed remedy is deliberately structural. Banning or restricting personalized advertising and data trading, minimizing collection, imposing duties on data custodians, improving cybersecurity, and deleting obsolete information address incentives rather than asking users to read every policy notice. Individual habits matter, but cannot substitute for public rules. A tension remains: personal data can support valuable research and services, while broad restrictions may make some beneficial uses harder. The book’s best answer is procedural rather than absolute: establish necessity, consent, safeguards, accountability, and public benefit before collection becomes routine.
The argument is most persuasive when it distinguishes targeted, justified uses from indiscriminate commercial extraction. Its caution about claims of safety and innovation is also important, though evidence about surveillance effectiveness and emerging technologies varies by domain. Overall, the book offers a powerful governing principle: collect less, concentrate less power, and make institutions justify access to information about people rather than treating that access as the natural price of participation.
Review Summary
Privacy is Power receives generally positive reviews, with readers praising its insights into data privacy issues and the data economy. Many find it eye-opening and appreciate the practical advice provided. Some criticize the book for being repetitive or overly pessimistic about technology. Readers value the philosophical perspective on privacy and its importance for democracy. While some find the arguments compelling, others feel the book lacks depth or nuance. Overall, it's considered a good introduction to privacy concerns in the digital age.
People Also Read
Glossary
Surveillance capitalism
Profit from behavioral dataAn economic model in which companies collect and analyze people’s personal data to predict behavior and sell targeted access to people’s attention. In the book, it describes how advertising became tied to extensive tracking, and how that business model concentrated power in technology firms and enabled influence beyond ordinary product sales.
Data vultures
Collectors exploiting personal dataVéliz’s metaphor for companies and other actors that gather personal information, analyze it, and profit from its use or transfer. The term emphasizes that data may be extracted from ordinary activities, combined across sources, and used without meaningful benefit or control for the people described by it.
Toxic data
Data that creates lasting risksPersonal information that can harm individuals, institutions, or societies because it is sensitive, vulnerable to theft, and attractive to actors with conflicting interests. Véliz compares it to a hazardous substance: collecting and retaining more than necessary creates exposure, while breaches or later reuse can make the harm difficult or impossible to reverse.
Real-time bidding
Automated auctions for ad placementA process in which advertisers compete electronically for the opportunity to show an ad to a particular person as a webpage loads. The process can transmit personal information to multiple potential advertisers, sometimes before a user has meaningfully consented. Véliz argues that this infrastructure expands data circulation and enables personalized advertising.
Differential privacy
Protect statistics with controlled noiseA method for analyzing a dataset while limiting what can be inferred about any particular person. It introduces carefully calibrated randomness into data or results so that useful aggregate patterns remain available, but individual contributions are harder to identify. The technique can reduce privacy risks, though it is not appropriate for every task and must be implemented correctly.
Obfuscation
Disrupt tracking with misleading dataThe deliberate addition of ambiguous, confusing, or false information to interfere with surveillance and data collection. Véliz presents it as a possible response when people cannot simply refuse to provide information to a business. It is not a universal tactic: it may be inappropriate where institutions have a justified legal claim to accurate information.
Chilling effect
Self-censorship under surveillanceA reduction in speech, inquiry, or participation that occurs when people expect their behavior to be monitored and potentially used against them. Véliz cites research finding a sharp drop in some Wikipedia searches related to terrorism after revelations of government surveillance. The concept explains how observation can constrain freedom even without direct punishment.
Fiduciary duties
Obligations to protect entrusted dataDuties of loyalty and care owed by an institution that is entrusted with something valuable and has power or expertise its customer lacks. Véliz proposes applying this idea to organizations holding personal data: they should prioritize the data subject’s interests, avoid conflicts of interest, and not exploit or disclose information in ways that harm the people who entrusted it.
FAQ
What's Privacy is Power about?
- Surveillance Capitalism Focus: The book explores how personal data is collected and traded by corporations and governments, leading to a surveillance society.
- Power Dynamics: Carissa Véliz argues that privacy is a form of power, and those who control personal data can dominate society.
- Call to Action: It provides insights on combating the surveillance economy and advocates for collective action to protect privacy rights.
Why should I read Privacy is Power?
- Understanding Data Vulnerability: The book offers a comprehensive overview of how personal data is exploited and the associated risks.
- Empowerment through Knowledge: Véliz aims to empower readers by providing the knowledge needed to reclaim their privacy.
- Practical Solutions: It highlights problems and offers actionable steps individuals can take to protect their data.
What are the key takeaways of Privacy is Power?
- Privacy Equals Power: Privacy is a form of power that enables individuals to maintain autonomy and resist manipulation.
- Toxic Nature of Data: Personal data is described as "toxic," capable of causing harm to individuals and society.
- Collective Responsibility: Privacy is a collective issue, requiring collective action to ensure a free and democratic society.
What are the best quotes from Privacy is Power and what do they mean?
- "Privacy is about being able to keep certain intimate things to yourself.": Highlights the importance of having private spaces in our lives.
- "The surveillance society has transformed citizens into users and data subjects.": Reflects the dehumanization that comes with surveillance capitalism.
- "Personal data is toxic, and we should regulate it as such.": Calls for strict regulations on personal data, likening it to hazardous materials.
How did we get to the current state of privacy according to Privacy is Power?
- Profit from Data: The data economy emerged as companies discovered the profitability of personal data.
- Post-9/11 Surveillance Expansion: Increased government surveillance under the guise of security post-9/11.
- Normalization of Surveillance: Society has become desensitized to privacy invasions, accepting them as the norm.
What is the surveillance economy as defined in Privacy is Power?
- Data Collection and Trade: A system where personal data is continuously collected, analyzed, and sold.
- Impact on Society: Undermines democracy and individual autonomy, creating power imbalances.
- Call for Reform: Advocates for dismantling the surveillance economy to protect civil liberties.
What practical steps does Privacy is Power recommend for individuals to protect their privacy?
- Limit Data Sharing: Be mindful of the data shared online and opt out of unnecessary data collection.
- Use Privacy Tools: Utilize tools like ad blockers and privacy-focused browsers to minimize tracking.
- Advocate for Change: Engage with policymakers and support organizations that fight for privacy rights.
How does Privacy is Power define the relationship between privacy and democracy?
- Privacy as a Foundation: Essential for a functioning democracy, allowing free expression without fear of surveillance.
- Power Dynamics: Control of personal data by corporations or governments threatens democratic processes.
- Collective Responsibility: Protecting privacy requires active participation from all citizens.
What is surveillance capitalism as defined in Privacy is Power?
- Definition: The commodification of personal data by tech companies for profit.
- Impact on Society: Leads to constant monitoring and manipulation, fostering distrust.
- Call for Regulation: Advocates for stronger regulations to curb these practices.
How does Privacy is Power suggest we improve cybersecurity standards?
- Government Regulation: Establish and enforce cybersecurity standards across industries.
- Collective Action Problem: A unified approach is necessary to protect everyone’s data.
- Designing for Security: Technology should be designed with security in mind from the outset.
What role do fiduciary duties play in the context of Privacy is Power?
- Prioritizing User Interests: Tech companies should prioritize user interests over profit.
- Accountability for Data Practices: Establishing fiduciary duties would hold companies accountable.
- Aligning Business Models: Aligning business models with user interests could create a more ethical data economy.
What is the significance of the right to be forgotten in Privacy is Power?
- Empowerment Through Forgetting: Allows individuals to request the deletion of their personal data.
- Addressing Past Mistakes: Essential for allowing individuals to move on from past mistakes.
- Balancing Public Interest: Must be balanced with public interest, not all data should be deleted.
Download PDF
Download EPUB
.epub digital book format is ideal for reading ebooks on phones, tablets, and e-readers.