Google has added a new way into locked accounts: a short selfie video that stands in for a password when the usual sign-in options fail. Announced on July 23, 2026, the feature, called "selfie video," is aimed squarely at account recovery, the moment a user is locked out because they lost a phone, forgot a password, or are stuck on a borrowed device. It joins passkeys, recovery emails, phone numbers, and recovery contacts rather than replacing any of them, and it lands at a time when AI-generated deepfakes have made identity verification a harder problem across the industry.<h3>How the Selfie Video Actually Works</h3><video src="https://cdn.calebms.com/articles/4d607e57-2838-4f3f-96f8-71da52bede6d.mp4" title="Google Selfie Sign-in" controls="true" style="width: 100%; height: auto; border-radius: 8px; margin: 1em 0px;"></video>Setup happens inside a Google Account's security settings, where users look into their device camera and complete several guided head movements so Google can capture their face from multiple angles. If they later get locked out, they record a fresh selfie video, and Google compares it against the enrolled version to confirm it is the same person. The head-turning requirement doubles as a liveness check, since real-time deepfake tools and face-swapping software tend to struggle with side profiles and sudden movement, making it harder to fool the system with a still photo or a pre-recorded clip. Users without a camera on their computer can hand off enrollment to a phone through a QR code.<ad-spot></ad-spot><h3>What Happens to the Video After You Record It</h3>Google says the video is encrypted at rest and used only to verify sign-in unless a user separately opts in to let it be used for other purposes, such as improving the company's facial recognition systems more broadly. That opt-in is off by default. Users can delete their enrolled video at any time from their account settings, and once deleted it can no longer be used for recovery. Google has noted it may keep deleted videos for a short period afterward for security purposes, and longer in cases tied to policy enforcement.<h3>Why Google Is Pushing This Now</h3>Account recovery has become one of the weakest points in modern authentication. Passkeys and hardware security keys have reduced how much people rely on memorized passwords, but the fallback path when someone loses their device still typically runs through a recovery phone number or email address, both of which attackers can intercept through SIM swaps or phishing. Google is framing the selfie video as a way to close that gap, especially since the accounts in question often hold years of emails, photos, and documents that would be difficult or impossible to replace.<ad-spot></ad-spot>The timing also lines up with a broader surge in AI-driven identity fraud. Industry research this year has pointed to a sharp rise in deepfake content and a growing number of publicly available face-swapping and voice-cloning tools, which has lowered the technical bar for impersonation attacks. That backdrop has pushed several major platforms toward biometric recovery methods that are harder to spoof than a static photo or a security question with a guessable answer.<h3>How This Compares to What Apple and Microsoft Already Offer</h3>Google is a late mover here rather than a pioneer. Apple has offered Face ID since the iPhone X launched in 2017, using it for device unlock, app authentication, and Apple Pay, though it is tied to Apple hardware rather than account-level recovery across devices. Microsoft's Windows Hello has supported facial recognition sign-in since Windows 10 launched in 2015, primarily as a way to unlock a specific device rather than to recover a locked account from anywhere. Google's version is different in scope: it works across devices and platforms as an account-level recovery method, closer to a biometric replacement for security questions than a device unlock feature.The feature is rolling out to eligible users now, with Google positioning it as a recovery-first tool even though it can also be used for regular sign-in. Its success will likely hinge on two things: how well the liveness checks hold up against increasingly capable deepfake tools, and whether users trust Google enough to hand over a biometric video of their face for something as high-stakes as account recovery. Security researchers have already flagged that any biometric system tied to something as permanent as a face carries higher stakes than a password, since a face cannot be reset the way a leaked password can. How Google handles edge cases, disputed identity checks, and the rare but real risk of a compromised enrollment video will be the real test of whether this becomes a trusted recovery option or another attack surface.

Google Now Lets You Recover Your Account With a Selfie Video

If you have ever tried to protect a piece of software in Kenya, you already know the drill. You register your code's copyright with the Kenya Copyright Board (KECOBO). If your product also involves a physical device, say an IoT sensor or a piece of hardware, you head over to the Kenya Industrial Property Institute (KIPI) for a patent or industrial design. And if someone starts selling counterfeit versions of your product, your next stop is the Anti-Counterfeit Authority (ACA). Three different agencies, three different portals, three different sets of forms, and three different institutional cultures, all for what is fundamentally one business problem: protecting what you built.That fragmented experience may be coming to an end. Parliament is currently reviewing a bill to merge KIPI, KECOBO, and the ACA into a single new agency, the Kenya Intellectual Property Authority. For Kenya's software developers, AI builders, and content creators, this is one of those quiet policy shifts that could actually change how they do business day to day.<h2>The Bill, In Plain Terms</h2>The proposal comes from a broader government push to shrink Kenya's sprawling list of state corporations. National Assembly Majority Leader Kimani Ichung'wah has tabled several bills seeking to dissolve or merge agencies with overlapping mandates, and the IP consolidation is one of four such mergers currently before the House. Under the plan, KIPI, KECOBO, and the ACA would be folded into one entity, the Kenya Intellectual Property Authority, with a single chief executive replacing the three currently in charge at the individual agencies.<ad-spot></ad-spot>This is not a brand new idea. The push to merge these three bodies goes back to a 2013 report by the Presidential Task Force on Parastatal Reforms, which found that the agencies performed similar and complementary functions and recommended housing them under one roof. That recommendation resurfaced in the Intellectual Property Bill of 2020, which proposed a similar merger under the name Intellectual Property Office of Kenya (IPOK). That version never quite made it across the finish line, but the idea persisted, and Cabinet gave its formal approval to the merger in January 2025. The current bill, tabled in July 2026, is the latest and arguably most concrete attempt to get it done, with Parliament having directed Treasury to complete this and other planned mergers by October this year.It is worth noting that the specifics can still shift as the bill moves through committee stages and public participation. What is fairly settled at this point is the direction of travel: Kenya is consolidating its IP administration into one authority.<h2>Why Fragmentation Has Been a Real Problem for Tech Builders</h2>If you are building software in Kenya today, your intellectual property rarely fits neatly into one legal box. A fintech app might involve source code (copyright, handled by KECOBO), a distinctive brand name and logo (trademark, handled by KIPI), and possibly a novel technical process worth patenting (also KIPI). An AI hardware startup might need patent protection for a device and copyright protection for the algorithms running on it, plus anti-counterfeiting support once a copycat product shows up on Luthuli Avenue or on an online marketplace.Each of these processes currently lives in its own silo. KECOBO handles copyright, and while registration is not mandatory since copyright protection in Kenya arises automatically once a work is fixed in tangible form, voluntary registration still matters in practice because it gives you a public record and stronger footing in a dispute. KIPI, on the other hand, handles trademarks and patents through a more formal examination process, and its own guidance suggests that straightforward trademark applications can take eight to eighteen months, with patent examinations sometimes stretching into years. The ACA operates on a completely different track altogether, focused on enforcement and seizure rather than registration.For a founder without an in-house legal team, that means learning three different bureaucratic systems, paying fees to three different agencies, and, in the worst case, discovering that inconsistent recordkeeping between agencies has left gaps that a counterfeiter or copycat can exploit. It is the kind of friction that rarely makes headlines but quietly taxes every startup that takes IP seriously.<h2>What a Single Authority Could Get Right</h2>The case for merging these agencies is genuinely strong, and it goes beyond just cutting a few CEO salaries. A single authority creates the real possibility of one account, one login, and one case management system across trademarks, patents, copyright, and counterfeit enforcement. If Kenya builds this thoughtfully, a developer could, in theory, register their app's brand, protect the underlying code, and flag counterfeit clones through one unified digital dashboard instead of juggling three logins and three sets of reference numbers.There is also a coordination benefit that matters more than it might seem at first glance. Enforcement bodies like the ACA are only as effective as the registration data they can draw on. When trademark and copyright records sit in separate systems, cross-checking a counterfeit claim against the original registration becomes slower and more error prone. A shared database changes that. It could mean faster, better substantiated enforcement action when a Kenyan developer's app gets cloned or when a hardware startup's design gets ripped off by an unlicensed manufacturer.<ad-spot></ad-spot>My honest take is that the upside here is less about the merger itself and more about whether it forces a genuine rebuild of the underlying technology. Kenya already has some digital infrastructure in place, KIPI runs an eCitizen-integrated portal for patents, trademarks, and industrial designs, and KECOBO has its own online systems for copyright. But these are separate systems built for separate agencies. Simply changing the letterhead on three websites and calling it a merger would be a missed opportunity. The real prize is a single, well-designed digital IP register built for how modern founders actually work, where a piece of software might need copyright, trademark, and eventually patent protection all within the same product lifecycle.<h2>Where It Could Go Wrong</h2>Mergers of this scale are never risk free, and Kenya's own history with parastatal consolidation gives reason for measured caution rather than blind optimism. The most immediate concern is transition disruption. Whenever government agencies restructure, there is typically a period where staff are reassigned, systems are migrated, and processing timelines that were already slow (patent examinations can take two to four years) risk stretching even longer as the new authority finds its footing. If you have a trademark or patent application in the pipeline when this transition happens, it is worth building in extra buffer time and keeping close tabs on which office is handling your file.There is also the practical question of institutional culture. KIPI, KECOBO, and the ACA have historically sat under different ministries and operated with different priorities: KIPI and the ACA under the Ministry of Investments, Trade and Industry, and KECOBO under the Attorney General's office. Merging their staff, systems, and case backlogs into one coherent authority is a genuinely difficult organisational task, and rushing it to hit a political deadline (Parliament has set October 2026 as a target for completing this round of mergers) could easily produce a bumpier ride than a more gradual transition would.<ad-spot></ad-spot>Finally, there is a legitimate worry among IP practitioners about added bureaucracy if the merger is not implemented cleanly. A bigger agency does not automatically mean a faster or simpler one. If the new authority inherits three separate approval chains instead of building one streamlined process, tech founders could end up facing the same fragmentation, just with a single logo stamped on top of it.<h2>What Tech Founders Should Do Right Now</h2>None of this is a reason to pause your IP strategy while waiting for Parliament to sort things out. If anything, the opposite is true. Here is what makes sense in the current window:<ul><li>File now rather than later. Copyright, trademark, and patent protections that you register today should carry over to the new authority once it is established, and getting ahead of any transition backlog is generally a good idea.</li><li>Keep your own records tidy. Since data migration between agencies is where things are most likely to get messy, keep your own copies of all registration certificates, application reference numbers, and correspondence.</li><li>Watch the public participation process. Bills like this typically go through committee review and public comment before final passage. Kenyan tech associations and startup hubs would do well to submit input on how the new authority's digital systems should actually work for software and AI products, rather than leaving that design entirely to legacy processes inherited from the merged agencies.</li><li>Do not skip registration just because copyright is "automatic." It is true that copyright protection in Kenya exists the moment original work is fixed in tangible form, but voluntary registration still gives you a much stronger position if you ever need to prove ownership in court or in a licensing negotiation.</li></ul><h2>The Bigger Picture</h2>This bill is really a bet that Kenya's IP administration works better as one coordinated system than as three separate fiefdoms. On paper, that bet makes sense, especially for an economy where software, content, and hardware innovation increasingly overlap in the same product. Whether it pays off will come down to execution: how well the new Kenya Intellectual Property Authority integrates its digital systems, how much of the current backlog it inherits, and whether the transition period is managed with the same urgency the government has applied to tabling the bill itself.For now, tech founders should treat this as a development worth watching closely rather than a reason to change their IP strategy today. The fundamentals of protecting your code, your brand, and your inventions in Kenya have not changed. What might change, hopefully for the better, is how much friction stands between you and that protection.

One Agency to Rule Them All: What the Kenya Intellectual Property Authority Bill Means for Tech Founders

Last week, we wrote about <a target="_blank" rel="noopener noreferrer nofollow" href="https://www.techinkenya.com/articles/google-breaking-20-year-promise-that-made-android-great">Google's Android Developer Verification Program</a> and how it represents a fundamental shift in what Android has always stood for. One of the natural questions that article raises is: if Google is tightening its grip on stock Android, what about the alternatives? What about GrapheneOS?It is a fair question, and on the surface the answer looks obvious. GrapheneOS is free, open-source, and built entirely outside of Google's corporate structure. It ships no Google apps by default, it cannot be controlled by Play Store policies, and it has consistently been rated as one of the most technically secure Android distributions available. If the concern is Google centralising control over the Android ecosystem, GrapheneOS looks like the logical refuge.Except that right now, in the same weeks that Google's verification program is dominating tech headlines, GrapheneOS users are getting locked out of their cars, flagged to law enforcement for choosing a private operating system, and treated as anomalies by platforms they pay to use. The escape hatch has its own wall, and it is being built by companies that are not even aware they are building it.<h2>What Is GrapheneOS, and Why Does It Exist</h2><img src="https://cdn.calebms.com/articles/1bbf8444-6c67-4efb-b095-a2b524bf1361.jpg" alt="GrapheneOS-y-Motorola.jpg" data-caption="">GrapheneOS is an open-source mobile operating system built on top of the Android Open Source Project (AOSP). It is developed by the GrapheneOS Foundation and targets users who want strong privacy and security without compromising on Android app compatibility. As of April 2026, the project has approximately 400,000 active users, a number that has grown significantly as privacy-conscious consumers have looked for alternatives to stock Android and iOS.The technical case for GrapheneOS is not subtle. It ships with a hardened memory allocator, more aggressive address space layout randomisation, stronger app isolation, and a smaller attack surface from Google services. Security updates ship faster than most OEM distributions. The project recently partnered with Motorola to bring official GrapheneOS support beyond Google Pixel devices, signalling that it is moving from a niche project to something more mainstream.<ad-spot></ad-spot>By every objective technical measure, a GrapheneOS device is more secure, more up-to-date, and more resistant to compromise than the average stock Android phone. That is not a point of controversy among security researchers.Which makes what is happening to its users right now particularly difficult to rationalise.<h2>Volkswagen: You Can Access Your Car From a Five-Year-Old Unpatched Phone, Just Not This One</h2>Volkswagen users running GrapheneOS began reporting recently that they could no longer log into the official VW app or remotely control their vehicles. The failure is not a bug. It is a deliberate enforcement decision.Volkswagen's app now requires a device to pass specific checks through Google's Play Integrity API, a system that verifies whether a device is running an officially certified Android environment. GrapheneOS does not pass those checks, not because it is insecure, but because it is not an official Google product and does not carry Google's proprietary certification. The result is that a Volkswagen owner running one of the most secure mobile operating systems available cannot unlock their own car from their phone.The comparison that GrapheneOS forum users are drawing is the one that matters most here. The VW app continues to work on stock Android devices that are completely end-of-life: phones that are no longer receiving security patches, carrying years of unaddressed vulnerabilities, potentially exposed to remote code execution exploits that have been public knowledge for years. Those phones pass. GrapheneOS, which ships timely security updates and is actively maintained, does not.<ad-spot></ad-spot>Volkswagen's response when users raised this was essentially a "sorry, unsupported" reply. The company confirmed that "alternative operating systems" including GrapheneOS, LineageOS, and similar custom ROMs are not part of the supported application environment for the VW app.That response is technically accurate and completely unsatisfying. Nobody is asking Volkswagen to provide technical support for custom ROMs. The question being raised is whether a car manufacturer should be using Google's certification framework as the sole definition of whether a device is trustworthy enough to control a vehicle, when that framework has no meaningful correlation with actual security.<h2>How Play Integrity Actually Works, and Why It Creates This Problem</h2>To understand why this keeps happening, it is worth being precise about what Play Integrity does and what it does not do.Google's Play Integrity API is a system developers use to verify three things: whether an app is genuine and unmodified, whether it came from the Play Store, and whether the device it is running on is trustworthy. That last check is the relevant one here.The API returns verdicts at different trust tiers. MEETS_BASIC_INTEGRITY means the device passes basic checks. MEETS_DEVICE_INTEGRITY means the device is a certified Android device with Google's approval. MEETS_STRONG_INTEGRITY is the highest tier, requiring hardware-backed security signals and, on Android 13 and above, a security update within the last year.GrapheneOS does not receive a MEETS_DEVICE_INTEGRITY verdict because certification requires devices to carry a cryptographic signature from a Google-approved OEM and include Google's proprietary components in a specific way. GrapheneOS is built entirely from open-source components and does not participate in that certification process by design.This is where the fundamental flaw lies. Developers integrating Play Integrity are not conducting a security audit of your device. They are checking a box: does Google vouch for this device? If yes, proceed. If no, block. The metric being used is compliance with Google's ecosystem, not actual device security. Those two things can overlap. They are not the same thing.<ad-spot></ad-spot>A five-year-old phone running a certified Android build passes MEETS_DEVICE_INTEGRITY. A fully patched, actively maintained, security-hardened GrapheneOS device does not. The API is measuring the right thing for detecting modified apps or rooted devices running sketchy software. It is measuring entirely the wrong thing when applied to a privacy-focused OS that is objectively more secure than what it is being compared against.Volkswagen's development team almost certainly did not sit down and decide to lock out GrapheneOS users. What they more likely did was turn on stricter integrity enforcement to block rooted phones and tampered app builds, and GrapheneOS became collateral damage. The end result is the same from the user's perspective, but the mechanism matters: this is not malice, it is outsourcing. Volkswagen handed their security model to Google's SDK and stopped thinking about it.<h2>Yoti: Using a Private OS Gets You Flagged to Police</h2>The Volkswagen situation is frustrating. The Yoti incident is something else entirely.Yoti is a British digital identity and age verification company whose tools are integrated into Sony PlayStation, TikTok, Meta, Spotify, and OnlyFans. Sony began rolling out mandatory age verification for PlayStation Network users in the UK and Ireland in 2026, driven by requirements under the UK Online Safety Act.A GrapheneOS user attempting to complete PlayStation's age verification through Yoti received a customer support response claiming their device had been "automatically reported to both the authorities and our security team." The stated reason: the user was running GrapheneOS, which Yoti cited as having "past security concerns."Yoti subsequently denied that the screenshots of the exchange match their records and suggested the messages may have been fabricated by the support agent. GrapheneOS publicly described the response as "customer support making ridiculous claims" and characterised it as fearmongering from an agent trying to close a ticket.Whether or not any actual law enforcement report was filed, the episode revealed something important about how GrapheneOS users are being treated by systems that rely on device fingerprinting for trust signals. When a platform cannot verify your device through Google's attestation framework, it does not conclude that you value privacy. It concludes that you are an anomaly. And in systems designed around fraud prevention, anomaly detection defaults to suspicion.The broader context around Yoti does not help their credibility here. In March 2026, Spain's data protection regulator AEPD fined Yoti approximately 950,000 euros for GDPR violations including improper biometric data retention, invalid consent mechanisms, and excessive data storage periods. Separately, researchers from the Georgia Institute of Technology and the University of California accused the company of collecting excessive data during age verification, including operating system version, RAM, and processor architecture, and sharing that data with third-party services including credit card companies, geolocation services, and data brokers. Yoti has denied those claims and called for the research to be withdrawn.A company with that specific track record flagging GrapheneOS users for "security concerns" is a particular kind of irony.<h2>The Pattern Being Established</h2>These are not isolated incidents. This is the third significant case in recent weeks where choosing a more private operating system has resulted in losing access to services. The consistent mechanism is the same in each case: Play Integrity certification, or a similar attestation framework, acts as a proxy for trust, and any device that does not participate in Google's approval chain gets flagged or blocked regardless of its actual security posture.The practical message being sent to privacy-conscious users is uncomfortable but clear: the more seriously you take your own security, the more mainstream services will treat you as a problem.This matters not just for the roughly 400,000 GrapheneOS users directly affected, but for what it signals about the direction of the mobile ecosystem more broadly. If Play Integrity becomes the de facto standard by which every significant app decides who gets access, then Google's certification framework becomes a quiet prerequisite for participating in digital life, even for users who have made a deliberate and technically informed choice to operate outside that framework.That is a form of control that does not require policy announcements or official mandates. It just requires enough developers to tick the "require device integrity" box without thinking too hard about what that box actually measures.<h2>GrapheneOS Is Not a Perfect Alternative Either</h2>This is a good moment to be honest about the limits of the GrapheneOS escape hatch, even before the Play Integrity problem is considered.GrapheneOS only runs on Google Pixel devices, with Motorola support beginning rollout in 2027. If you do not own a Pixel, you cannot run it. The installation process requires comfort with flashing firmware and working through a command-line web installer: it is more accessible than it used to be, but it is still not a mainstream experience. There is no equivalent of walking into a phone shop and asking for a GrapheneOS handset.Beyond the technical barrier, the compatibility picture has always required compromise. Many popular apps either do not work or require workarounds involving GrapheneOS's sandboxed Google Play feature, which allows running Google's app framework in a restricted environment without giving it system-level access. That sandboxed approach works well for most apps, but it cannot spoof Play Integrity attestation at the highest device integrity tiers, which is exactly why the Volkswagen and Yoti situations are happening.<ad-spot></ad-spot>The project has also faced its own institutional pressures. GrapheneOS ceased operations in France in late 2025 following government pressure and legal threats. The project publicly stated in March 2026 that it will not comply with legislation requiring collection of user age data, which reflects a principled stance but also signals ongoing friction with regulatory environments in multiple jurisdictions.None of this makes GrapheneOS a bad option. For the users it serves, it remains technically exceptional. But it is not a seamless escape from the trade-offs of the mainstream mobile ecosystem. It trades one set of constraints for another.<h2>The Actual Problem, Plainly Stated</h2>The root issue is not Google, or Volkswagen, or Yoti specifically. The root issue is that device attestation has become infrastructure, and that infrastructure has a single controlling authority.When Google's Play Integrity API becomes the standard mechanism by which apps decide whether a device is trustworthy, Google's definition of "trustworthy" becomes the effective policy for the entire ecosystem. That definition, by design, excludes any operating system that does not carry Google's certification. Not because those systems are insecure, but because Google cannot commercially certify what it does not control.Companies like Volkswagen are not running security audits. They are delegating the security decision to Google's SDK and moving on. That is understandable from an engineering resource perspective. It is also how a monopoly on trust gets built without anyone explicitly choosing to build one.For users, the ask should be simple: if your app supports remotely accessing a vehicle, controlling financial accounts, or verifying someone's identity, you should be able to explain your security model in terms of what the device actually does, not just whether Google has approved it. An operating system that ships security patches faster than your stock Android OEM does, that isolates apps more aggressively, and that does not share device telemetry with advertising ecosystems should not be disqualified from "trusted" status because it chose not to carry Google's paperwork.The Android Developer Verification Program is changing how apps get distributed on Android. The Play Integrity API is quietly shaping who gets to use them at all. Both trends are worth watching, and right now they are pointing in the same direction.

The Monopoly on Trust: How Google Is Quietly Locking You Out of Your Own Hardware

You open an article on Fortune or Business Daily. A paywall modal slides up. You are told to subscribe for access to the full story. But if you open your browser's developer tools and look at the Network tab, the full article is already sitting in the response (every paragraph, every sentence ) delivered to your browser before the subscription prompt even appeared.The paywall is not protecting the content. It is hiding it. Those are not the same thing.This is not a new discovery and it is not a niche technical quirk. It is a deliberate architectural choice made by some of the largest media companies in the world and understanding why they made it reveals something genuinely interesting about the tension between subscription revenue, search engine visibility, and the fundamental principle of web security that every developer learns early: never trust the client.<ad-spot></ad-spot><h2>Two Fundamentally Different Architectures</h2>There are two ways to build a paywall. They produce similar-looking results for the end user but are architecturally opposite in how they handle the content.<h3>Client-Side Paywalls</h3>When you visit a paywalled article on a site using a client-side paywall, the server sends you everything. The full article HTML arrives in your browser. JavaScript then runs, checks whether you have an active subscription, and if not, renders an overlay ( a modal, a blur, a truncation ) that obscures the content from view.The content is already in the browser. The paywall is presentation logic, not access control. It is a curtain drawn over something that is already in the room.The practical consequence: anything that interrupts that JavaScript execution reveals the content. Disabling JavaScript entirely. Using a browser's built-in reader mode. Using a read-it-later extension that strips page chrome and renders only the article text. Intercepting the network response before JavaScript processes it. None of these require technical expertise. They are features built into mainstream browsers.<ad-spot></ad-spot>Firefox's reader mode ( accessible via the book icon in the address bar or pressing F9 ) is the simplest example. It parses the article content directly from the HTML and presents it in a clean, ad-free, popup-free reading view. On a client-side paywall, this works because the content was already delivered to the browser. Reader mode simply presents what is there without executing the JavaScript that was supposed to hide it.<h3>Server-Side Paywalls</h3>Bloomberg is the clearest example of the opposite approach. When you hit a Bloomberg article without a subscription, the server does not send you the full article. It sends you the headline, the first paragraph, and a subscription prompt and that is all that exists in the HTTP response. There is nothing else in the page source to find.No reader mode. No JavaScript manipulation. No network interception. No amount of browser tooling reveals the rest of the article because the rest of the article was never sent. The access control decision was made on the server, before a single byte of article content was transmitted.The difference in developer terms is simple:Client-side: The server sends content → the browser decides who sees itServer-side: The server decides who gets content → then sends itThe second model correctly treats authentication and authorisation as a backend concern. The first model treats it as a frontend concern, which any developer who has built a web application knows is the wrong layer for access control.<h2>So Why Would Fortune Do This?</h2>Fortune is not a company of naive engineers. Its digital infrastructure is backed by significant investment and experienced teams. The choice to implement a client-side paywall is not an oversight. It is a calculated trade-off, and the primary currency in that trade-off is SEO.Here is the problem Bloomberg faces with its server-side paywall: Google cannot read the article either. When Googlebot crawls a Bloomberg page, it sees the same truncated content a non-subscriber sees, the headline and the first paragraph. Google cannot index the full text of the article. The article cannot rank on long-tail search queries for its specific content. Bloomberg accepts lower search visibility in exchange for a genuinely enforced paywall.Fortune, and publications like it, make the opposite trade. By delivering the full article to every browser ( including Googlebot ) the full text is indexed by Google. The article ranks for every keyword in the body of the piece, not just the headline. Readers who find it via search land on the full content, hit the paywall overlay, and are invited to subscribe. The content does the SEO work before the paywall does the conversion work.<ad-spot></ad-spot>Client-side metered paywalls allow the full HTML content to be indexed by search engines before any JavaScript-based restrictions are applied. This approach is preferred for SEO as it ensures that search engines can fully crawl and index the content, which is vital for maintaining search rankings.This is a known and accepted SEO strategy, not a secret. Google has explicit guidance for publishers doing this, they are required to use structured data to declare that the content is paywalled, specifically the <code>isAccessibleForFree: false</code> attribute in NewsArticle schema markup. This prevents the practice from being classified as cloaking ( showing different content to search bots than to users ) which would be a Google Search penalty.Whether every publication implementing client-side paywalls correctly implements this structured data disclosure is a separate question.<h2>The SEO vs Revenue Trade-Off in Plain Terms</h2>The choice between client-side and server-side paywalls is ultimately a question of which you value more: content protection or content discovery.Server-side paywall (Bloomberg model):<ul><li>Content is genuinely protected, no bypass possible</li><li>Search engines cannot index full content</li><li>Relies on brand authority and direct traffic for subscriber acquisition</li><li>Works for publications with strong established brands where readers seek them out directly</li></ul>Client-side paywall (Fortune model):<ul><li>Content is technically accessible via reader mode and JS manipulation</li><li>Search engines index full content, driving organic discovery</li><li>Works for publications that depend on search traffic to feed the subscription funnel</li><li>Acceptable leakage rate (the proportion of technically capable readers who bypass the paywall) is judged to be smaller than the SEO benefit</li></ul>The Business Daily Kenya context is instructive here. A publication in a market where search discovery is critical for audience growth (and where a Bloomberg-style brand authority that drives direct traffic does not yet exist ) has strong incentive to prioritise search indexing over airtight content protection. The readers most likely to bypass the paywall are the least likely to convert to subscribers anyway. The readers who find the article through Google search and see a subscription prompt are exactly the conversion target.<h2>Reader Mode and the UX Problem It Solves</h2>Beyond the paywall question, browser reader modes are solving a separate and legitimate problem: modern news websites have become hostile environments for reading.<ad-spot></ad-spot>Autoplay videos that play on mute. Newsletter signup modals that appear mid-scroll. Cookie consent banners that obscure half the page. Sticky headers that eat 15% of the viewport. Related article carousels that interrupt the reading flow. Push notification requests. Live chat widgets. None of these serve the reader's primary intent, which is to read the article.Reader mode strips all of this away. The result is a page with the article text, the images relevant to the story, and nothing else. Loading times drop dramatically. Battery consumption falls. Mobile data usage decreases.The irony is that reader mode produces a better reading experience than most publishers' own article pages. The sites that invest most heavily in ad units, promotional widgets, and engagement popups (the ones most motivated to drive subscription revenue ) have often built reading environments that actively discourage reading.Firefox's built-in reader mode preserves the article images but strips the surrounding page furniture. Extensions like Ream go further, rendering images inline with the article text in a fully typeset layout. For long-form articles on a phone commute, the difference in experience is significant.<h2>What This Means if You Are Building a Paywall</h2>If you are a developer building subscription access control into a web application ( whether that is a media publication, a software tool, or any content platform ) the architectural lesson is straightforward.Access control belongs on the server. The browser is not a trust boundary. JavaScript can be disabled, intercepted, modified, or simply not executed. Any access control decision that runs exclusively in the browser is a presentation decision, not a security decision. The distinction matters: presentation logic controls what users see, security logic controls what data they receive.The correct pattern for any content that genuinely needs to be restricted:1. The client requests the content2. The server checks authentication and authorisation3. If the user is not entitled to the content, the server returns the teaser or a 401/403 response, not the full content4. If the user is entitled, the server returns the full contentThis is not complicated. It is the same pattern used for every other authenticated resource on the web, private API endpoints, protected file downloads, user account data. The reason paywalls often deviate from it is the SEO trade-off described above, not a technical limitation.If you are building a paywall and you want both SEO indexing and content protection, the technically correct approach is to implement server-side access control for human users while explicitly allowing known Googlebot user-agents to receive full content, a practice Google permits and documents, provided you implement the correct structured data declarations. This is more complex to build than either pure approach but eliminates the trade-off.<h2>The Broader Principle: Never Trust the Client</h2>The paywall case is a highly visible instance of a principle that applies across every layer of web development.Frontend validation of form inputs is not a substitute for backend validation ( a user can submit any data directly to your API endpoint regardless of what your form allows. Client-side role checks in a React app are not access control) anyone can modify the component state or call the API directly. A disabled button is not a security boundary. CSS that hides content is not content protection.<ad-spot></ad-spot>These are not obscure security concepts. They are in every web development curriculum, every security awareness training, and every code review checklist. The reason client-side paywalls exist despite this is not ignorance, it is a deliberate product decision where search visibility was judged more valuable than content protection.That judgement may be correct for the publications that made it. But it is a product decision, not a technical best practice. Knowing the difference, and knowing why a Fortune-sized company makes that choice, is what separates engineers who implement features from engineers who understand systems.Are you building a paywall or subscription access control for a web application? Drop your architecture questions in the comments, we are happy to dig into the implementation details.

Client-Side vs Server-Side Paywalls: Why Some of the World's Biggest Media Companies Are Trusting the Browser With Their Revenue

Between December 2025 and January 2026, <a target="_blank" rel="noopener noreferrer nofollow" href="https://www.bloomberg.com/news/articles/2026-02-25/hacker-used-anthropic-s-claude-to-steal-sensitive-mexican-data">a single unknown person used a consumer AI chatbot to systematically dismantle the cybersecurity defences of the government of Mexico</a>. By the time anyone noticed, 150 gigabytes of data was gone — tax records tied to 195 million taxpayers, voter files, government credentials, and civil registry documents spanning Mexico's federal tax authority, its electoral institute, four state governments, and a water utility.The tool was not a sophisticated nation-state hacking suite. It was a consumer AI chatbot. The attacker's conversation logs were found publicly accessible online when the breach came to light on February 25.The method was not technically novel. What made it effective was patience and AI-assisted persuasion, the ability to craft perfectly targeted social engineering at scale, probe defences continuously without fatigue, and adapt to each new barrier without human oversight slowing the attack down.Kenya has built something similar to what Mexico lost except larger and more centralised. Over 13 million Kenyans are registered on eCitizen. More than 16,000 government services sit on a single platform. The Maisha Namba digital ID is issuing up to 30,000 new IDs every day. The question of whether Kenya's cyber defences are ready for AI-assisted attacks is not hypothetical. It is urgent.<h2>The Ksh 29.9 Billion We Already Know About</h2>Before discussing where cyber threats are heading, the current baseline deserves to be stated clearly.Kenya lost Ksh 29.9 billion to cybercrime in 2024/2025. That figure comes from the <a target="_blank" rel="noopener noreferrer nofollow" href="https://www.serianu.com/downloads/Africa%20Cybersecurity%20Report%20-%20Kenya.pdf">Africa Cybersecurity Report Kenya 2024/2025,</a> produced by Serianu, one of the continent's most credible cybersecurity research institutions, based on responses from 280 organisations. Continent-wide, Africa lost Ksh 650 billion, roughly $5 billion, in the same period.<ad-spot></ad-spot>Payment fraud was the most common incident category, driven by real-time mobile money transfers, weak transaction monitoring systems, and social engineering attacks that exploit the trust that makes M-Pesa so frictionless in legitimate use. Online and email fraud accounted for 40% of all incidents and 32% of recorded losses, pointing to persistent gaps in how organisations manage digital identity verification. SIM-swap attacks and mobile money fraud remain widespread despite what Serianu describes as "improved controls", a diplomatic way of saying the controls are not keeping up with the attackers.A specific case from 2025 illustrates the gap between controls and outcomes: attackers compromised a digital payments portal, disabled OTP notifications, and funnelled Ksh 49 million into mobile wallets, bank accounts, and till numbers before the breach was detected. In a separate incident, a syndicate stole over Ksh 6 million from a commercial bank. Both were carried out with relatively conventional methods. Neither required AI.The Ksh 29.9 billion figure is sobering on its own. What Serianu's CEO William Makatiani said at the report's launch is more sobering still: "The rate at which organisations are investing is improving, but attackers are advancing even faster."That gap, between the pace of investment in defence and the pace of offensive capability, is where AI enters the picture.<h2>What AI Has Already Changed for Attackers</h2><a target="_blank" rel="noopener noreferrer nofollow" href="https://www.bcg.com/publications/2025/ai-raising-stakes-in-cybersecurity">A global survey of 500 senior leaders</a> by Boston Consulting Group, published in December 2025, quantified the defensive side of this gap. Only 7% of organisations globally have deployed AI tools in their cyber defence. Meanwhile, approximately 60% have likely already experienced an AI-powered attack in the past year. The defensive majority is fighting with 2019-era tools against attackers who have fully adopted 2026 capabilities.<ad-spot></ad-spot>The BCG report identifies three structural shifts that make AI-enabled attacks categorically different from what came before.The first is scale and speed. AI enables attackers to hunt for vulnerabilities across thousands of systems simultaneously, generate personalised phishing content at industrial volume, and probe defences continuously without the fatigue or cost constraints that limit human attackers. A campaign that previously required a team of analysts working over weeks can now be initiated and iterated by one person with a laptop and an API key.The second is autonomous operation. AI has moved from static models to agentic systems, AI that can observe, reason, and act across multiple steps without human oversight. The Mexico breach demonstrated this: the attacker used AI to sustain a patient, multi-stage intrusion across months, adapting to each defensive response without escalating resources or personnel. Autonomous attacks can probe continuously, adapt instantly to defensive signals, and maintain operational tempo that human defenders cannot match in real time.The third is identity collapse. Deepfakes, voice cloning, and synthetic identity generation have undermined authentication as a reliable safeguard. The BCG survey found that AI-generated deepfake video has already been used to impersonate a CFO in a live video call, resulting in a $25 million fraud loss at a multinational engineering firm. Voice cloning was used to spoof election-related robocalls, resulting in a $1 million regulatory fine for a telecom provider. When you cannot trust that the voice or face on your screen belongs to the person it claims to be, the entire model of identity-based verification collapses.Kenya's digital economy is built on identity verification. M-Pesa PIN entry. OTP confirmation. Biometric matching at eCitizen kiosks. SIM registration against national ID. Every one of these controls was designed for a threat environment where forgery was expensive and required physical proximity. AI has made sophisticated forgery cheap, scalable, and remote.<h2>Why Kenya Is a High-Value Target</h2>Kenya's appeal to cybercriminals is a function of the same qualities that make its digital economy a regional success story.M-Pesa processes hundreds of billions of shillings in transactions every month, with the frictionless speed that makes mobile money genuinely useful also creating opportunities for real-time fraud before monitoring systems flag anomalies. The eCitizen platform centralises access to government services at a scale that, if compromised, would expose millions of citizen records in a single breach. Kenya's position as East Africa's technology hub means it hosts the regional offices, data centres, and payment processing infrastructure of companies operating across multiple countries, making it a gateway target whose breach has continental consequences.<ad-spot></ad-spot>The Serianu report notes that Kenya's digital infrastructure remains significantly exposed at the network layer. Many devices remain accessible through open, unencrypted remote-access services (Telnet, FTP, RDP) that should have been secured or decommissioned years ago. Operational errors and system misconfigurations compound the attack surface, particularly in public sector institutions where IT staffing and budget constraints are most acute.Kenya also has the third-highest number of internet users in sub-Saharan Africa and among the highest mobile money penetration rates globally. From an attacker's perspective, this is not a liability, it is an opportunity. More users, more transactions, more endpoints, more attack surface, and a regulatory and enforcement environment that is still building the capacity to respond at the speed and sophistication that AI-enabled attacks require.<h2>The eCitizen Problem Nobody Is Talking About Loudly Enough</h2>The Mexico breach warrants specific attention in the Kenyan context because the structural parallels are uncomfortable.Mexico's breach exploited centralised government infrastructure, a single point of failure whose compromise gave the attacker access to data spanning multiple agencies. eCitizen is Kenya's equivalent architecture. The platform's value proposition is precisely its centralisation: one login, one identity, access to thousands of services. That centralisation is also its vulnerability. A successful attack against eCitizen's authentication layer, or against the identity verification systems that underpin Maisha Namba, would not be a single agency breach. It would be a systemic one.Kenya's government moved faster on digital transformation than on the security architecture to protect it. The Maisha Namba rollout, the eCitizen expansion, the integration of biometric data across health, education, and financial services, these are genuinely impressive infrastructure achievements. But each integration point is also a potential attack surface, and the BCG survey's finding that 88% of organisations plan to deploy AI defence but only 7% have done so yet describes a gap that attackers are exploiting right now.The government is aware of this. Secretary to the Cabinet Mercy Wanjau, speaking at the Serianu report launch, announced the establishment of a Government Security Operations Centre, a centralised national cyber-monitoring facility. That is the right structural response. The implementation timeline, resourcing, and integration with existing agency security teams will determine whether it arrives before the next major incident or after it.<h2>What Organisations and Individuals Can Do Now</h2>For businesses and institutions operating in Kenya, the BCG and Serianu findings point to concrete priorities rather than abstract awareness.Identity governance is the most urgent gap. Both reports identify identity management, how you verify who is accessing what, as the primary attack surface for AI-enabled threats. Multi-factor authentication that relies solely on SMS OTPs is no longer sufficient, as SIM-swap attacks can intercept those codes. Authenticator apps, hardware security keys, and biometric verification tied to the actual device rather than a phone number are more resilient options. For organisations processing significant financial transactions, these are not optional upgrades.<ad-spot></ad-spot>Recovery readiness matters as much as prevention. The shift Serianu recommends (from risk management to resilience) reflects a hard-won insight from organisations that have been breached: prevention will sometimes fail. Immutable backups, routine recovery testing, and documented incident response procedures determine whether a breach becomes a recoverable disruption or a catastrophic loss. The organisations that recovered fastest from the attacks documented in the Serianu report were those with functioning recovery systems, not necessarily those with the most sophisticated perimeter defences.AI-assisted monitoring is now table-stakes for financial institutions. Real-time transaction monitoring that flags anomalies (unusual patterns, atypical access times, out-of-pattern transfer destinations) needs to operate at the speed of AI-enabled attacks, which means it needs AI assistance. Human analysts reviewing logs after the fact cannot catch attacks that complete in minutes. This is not a future consideration for Kenya's banks, fintechs, and mobile money operators. It is a current gap that Serianu's documented incidents have already exploited.Consumer awareness remains the cheapest layer of defence. Enabling two-step verification on all financial accounts. Confirming transfer requests through a separate channel before executing them. Being sceptical of urgent requests from known contacts (particularly voice or video calls requesting money or credentials) since both can now be convincingly faked. These behaviours cost nothing and intercept the social engineering attacks that remain the most common entry point for Kenya's most costly breaches.<h2>The Asymmetry That Needs to Close</h2>Kenya is in a position that is not unique but is particularly acute. It has built digital infrastructure fast enough to create real economic value (M-Pesa's role in financial inclusion, eCitizen's role in service delivery, Maisha Namba's potential to anchor a unified digital identity) but the security architecture protecting that infrastructure has not kept pace with either the sophistication of the threats or the centralisation of the target.The Ksh 29.9 billion in losses documented by Serianu is a 2024/2025 figure. It reflects attacks carried out with largely conventional tools. AI-enabled attacks are already underway globally, the BCG survey's 60% figure represents organisations that have already experienced them, and Kenya's attack surface is large, high-value, and in several areas still inadequately protected.The Mexico breach should be read as a preview, not a warning about somewhere else. The same consumer AI tools are available to the same global threat actors targeting Kenya. The centralised digital infrastructure that makes Kenya's government services efficient makes it a similarly attractive target. The response gap, between organisations that understand the risk and those actively deploying AI-native defences, is the window that attackers are operating in right now.Closing that window is not primarily a technology problem. It is a resource, urgency, and prioritisation problem. The technology to defend against AI-enabled attacks exists. The organisations and institutions that deploy it at speed and scale will be the ones that are not in next year's Serianu report.

Google Now Lets You Recover Your Account With a Selfie Video

One Agency to Rule Them All: What the Kenya Intellectual Property Authority Bill Means for Tech Founders

The Monopoly on Trust: How Google Is Quietly Locking You Out of Your Own Hardware

Client-Side vs Server-Side Paywalls: Why Some of the World's Biggest Media Companies Are Trusting the Browser With Their Revenue

Kenya Lost Ksh 29.9 Billion to Cybercrime Last Year. AI Is About to Make That Number Much Worse.