We Are Building the Safest Place on the Internet. That Requires Saying What We Will Not Allow

Most technology companies talk about safety in the passive voice. Harm “occurs.” Content “appears.” Bad actors “exploit” the platform, as though the platform were weather rather than a set of deliberate engineering decisions made by named people who could have decided otherwise.

LINKSPREED is going to speak in the active voice, because the record of the last twenty years shows what passive-voice safety produces.

What the record actually shows

In 2025, the Internet Watch Foundation assessed 451,210 reports. 311,610 were confirmed to contain or lead to child sexual abuse material — one report assessed every 70 seconds, one confirmed case every 101 seconds, with reports rising roughly 6% year on year. A single report can contain tens of thousands of images.

The composition of that data should end any remaining debate about whether online abuse is a marginal phenomenon.

It is overwhelmingly gendered. 77% of victims in non-AI-generated material are girls. In AI-generated material, that figure is 97%. IWF analysts describe material depicting girls as carrying misogynistic and violent tones.

It is being industrialised by generative AI. AI-generated video containing child sexual abuse rose 260-fold in 2025, reaching 3,443 videos; across two years, 15,036 AI-generated images and videos have been identified. 65% of the 2025 AI-generated video material was Category A — the most severe classification — compared with 43% where AI was not a factor. Generative tools have collapsed the technical barrier to producing this material, and because real victims’ images are used to train the models, new imagery of known victims is continually created, re-victimising children who were abused years ago.

It is commercial, and growing fast. Commercial child sexual abuse websites identified by IWF analysts rose from 7,028 in 2024 to 15,031 in 2025 — more than doubling in one year. Many were engineered to appear non-criminal or dormant so they could operate on the open web while evading detection. Many accepted cryptocurrency, card payments, or money transfer services.

It is extorting teenagers at scale. Reports from young people experiencing sexually coerced extortion rose 127% compared to 2024. In financially motivated sextortion, boys aged 14–17 represent 98% of victims. The IWF documents how fast it escalates — more content, more explicit content, more money — and notes organised criminal groups deliberately targeting children. In some cases, the images were captured without the child’s knowledge at all.

Why the existing internet failed these children

The Web2 failure was a failure of incentive. When a platform’s revenue is a function of engagement, safety is a cost centre, and cost centres get managed downward. The consequences are now in the public record through litigation. Instagram’s former Head of Safety and Well-being, Vaishnavi Jayakumar, testified that accounts known to engage in trafficking operated under a 17-strike policy — sixteen violations for prostitution and sexual solicitation before suspension on the seventeenth — a threshold she described as extraordinarily high by any industry measure. Her testimony further described a “zero tolerance” policy for child sexual abuse material that was a policy in name only, on a platform that did not offer users a simple way to report such content, an issue she raised repeatedly from 2020 onward.

That is not a technical failure. A company capable of ranking billions of pieces of content per second is capable of building a report button. It is a failure of priority, and priority is a choice.

The Web3 failure was a failure of architecture. The pitch was that removing identity removes control, and removing control produces freedom. What it actually produces is a system in which no one can be excluded, because no one can be identified.

The consequences are measurable. Illicit cryptocurrency addresses received at least $154 billion in 2025, a 162% year-on-year rise. Money launderers received at least $82 billion. Flows to suspected human trafficking services rose 85% year over year. Darknet markets, after a decade of international takedowns, still processed slightly more than $2.5 billion in 2025 — more than the year before. And commercial abuse sites accept cryptocurrency among their payment rails.

Pseudonymity was sold as protection for dissidents. In practice, it has functioned as protection for whoever most needs not to be found — and the people who most need not to be found are rarely dissidents.

Our position: ownership and accountability are the same technology

Here is the argument the last two decades got wrong, and the one LINKSPREED is built on.

The industry has treated privacy and accountability as opposites — as though you must choose between a surveillance economy that knows everything about you and sells it, or an anonymity economy that knows nothing about anyone and therefore protects predators. Both choices are bad, and both are false.

Verified identity and genuine privacy are not in tension. They are produced by the same cryptography.

UIID — Universal Integrated Identity, Decoupled — is LINKSPREED’s self-sovereign identity system: secure, cryptographically protected identity that the user controls rather than a corporation. It functions as a decentralised digital passport that decouples identity from any single platform and makes the individual the ultimate authority over their own data.

Concretely, here is how it resolves what looked like a contradiction:

You are private from platforms. The dual-layer architecture splits your presence into a master Core ID and context-specific aliases, structurally preventing services from cross-correlating your behaviour across the web. Bridge a legacy Web2 login into a sovereign key, and the connection to legacy tracking mechanisms is permanently severed. No behavioural profile. No data farming for ads.

You are not anonymous to the law. Digital identities are anchored in hardware using WebAuthn standards and localised biometrics — which is precisely what makes throwaway profiles, deepfake accounts and mass bot networks economically and technologically impractical. UIID additionally provides instant verification for bot deterrence.

This is the asymmetry we consider correct and non-negotiable: maximum privacy against commercial surveillance, zero refuge from lawful investigation of serious crime.

Web4 is not an anonymous network. You need a verified identity to enter it. That is not a limitation we apologise for. It is the entire security model. An internet where anyone can be nobody is an internet optimised for the people with the strongest reasons to be nobody.

What we will not allow, stated plainly

We are being specific because vagueness is what predators navigate by. The following are strictly prohibited on any LINKSPREED-operated Web4 infrastructure, and will be treated as matters for law enforcement rather than content moderation:

• Child sexual abuse material, in any form — photographic, non-photographic, or AI-generated. UK law treats realistic AI-generated material identically to photographs of real children, and so do we.
• Grooming, solicitation, and the sexual exploitation of minors, including the coordination or advertising of access to children.
• Sexually coerced extortion of any person, adult or child — the fastest-growing threat to teenagers on the internet today.
• Human trafficking and the recruitment, advertisement, transport, or financing of it.
• Terrorism, its financing, and violent extremist organisation.
• Trafficking in controlled substances, weapons, and stolen identity or payment data.
• Money laundering, sanctions evasion, and the infrastructure that services them.
• Ransomware, malware distribution, and coordinated intrusion.
• Non-consensual intimate imagery, including synthetic and AI-generated depictions of real people.

How accountability works — precisely

We are going to describe this exactly, because both overpromising and underpromising cost lives.

LINKSPREED is not a police force, and we will not pretend to be one. Private vigilantism contaminates evidence, endangers victims, and collapses prosecutions. What we operate instead is identity-bound accountability, which is considerably more dangerous to offenders than any amount of rhetoric:

Entry requires verified identity. Hardware-anchored, biometrically bound, cryptographically issued. There is no anonymous door. There is no disposable account. The economics of throwaway identity — the foundation of every abuse operation on the open internet — do not function here.

Detection is proactive and standards-based. We align with the established global infrastructure for this work. The IWF operates Europe’s largest centre for assessing, grading and securing removal of child sexual abuse material, providing trusted services, intelligence and datasets used by technology companies, financial institutions and law enforcement worldwide — including hash lists and URL lists designed for exactly this purpose. Safety-by-design is not a press release; it is a set of specific, available technical integrations, and we take them.

Evidence is preserved to an evidentiary standard. When criminal conduct is identified, the account’s cryptographic identity anchors, access records and associated material are preserved intact for lawful disclosure — not scrubbed, not silently deleted, not quietly deplatformed in a way that destroys a prosecution.

Referral is immediate, and to the proper authority. We report to national law enforcement and the relevant hotline and reporting bodies in the offender’s jurisdiction, and we cooperate fully with valid legal process.

There is no safe harbour. Because identity is hardware-anchored and verified at entry, an offender on Web4 infrastructure is not a pseudonym, a wallet address, or an IP behind seven proxies. They are a person, and they are identifiable to the authorities of the country whose laws they broke.

If you come to Web4 to harm children, you will not find the thing you have relied on everywhere else. You will not find the crowd to hide in.

An open invitation to governments

We want to say this without hedging: LINKSPREED wants to work with governments, regulators, and law enforcement agencies.

That statement is unusual in this industry, and we understand why. A generation of technology companies has treated regulators as an adversary to be lobbied, litigated, and outlasted. We regard that posture as one of the primary reasons the figures in this article exist.

The IWF’s own conclusion is that tackling online child sexual abuse requires a collaborative approach between industry, government, and foundation partners, and that this is paramount. We agree completely, and we are prepared to be a constructive partner rather than a reluctant one.

We are open to work with authorities on verified-identity standards that satisfy both child protection duties and data protection law; on lawful-process interfaces that are fast, auditable, and resistant to abuse; on hash-sharing and detection integration with established hotlines; on cross-border evidentiary cooperation; and on independent audit of our safety systems. We would rather have our safety architecture examined and criticised by serious people than praised by nobody.

We are equally clear about the boundary. Cooperation with lawful investigation of serious crime is not the same as building surveillance infrastructure for political or commercial purposes, and we will not confuse the two. The architecture is deliberately built so that these remain distinct: platforms cannot correlate your behaviour, advertisers cannot profile you, and none of that changes when a court issues a valid order regarding a specific serious crime.

Why we are doing this

Because 311,610 confirmed reports in one year is not a statistic. It is a number of children.

Because a 127% rise in sexually coerced extortion reports means a specific fourteen-year-old, somewhere, being told tonight that the images will be sent to everyone he knows unless he pays.

Because a 260-fold increase in AI-generated abuse video means the worst people on earth have been handed an industrial tool, and the response cannot be another statement about how seriously we take these issues.

LINKSPREED was founded to build an internet where people own their identity, their data, and their communities. We have come to believe that ownership and protection are the same project. A person who does not own their identity cannot be protected — and a network where no one holds an identity at all cannot protect anyone.

We are building the safest place on the internet. Not the most permissive. Not the most anonymous. The safest — for the people who have spent two decades being told that their safety was a regrettable trade-off against someone else’s business model or someone else’s ideology.

That trade-off is over. Bring your identity. Bring your community. Bring your children.

Leave everything on the list above at the door — because here, unlike everywhere else, the door knows who you are.