For years, governments have searched for a way to make the internet safer for children. The latest answer is age verification, age estimation and social media bans. They sound logical. If children cannot access online services, surely they cannot be harmed.
Unfortunately, the world is not that simple.
A child can still receive an explicit image through a messaging app, a gaming platform, AirDrop or a livestream. They can still be groomed by another child. They can still be pressured into creating and sharing intimate images. None of these harms disappear simply because someone's age has been verified.
The problem is that age verification asks the wrong question.
It asks who should be allowed to access a service or create a social media account.
It does not ask how that child will be protected once they are using it.
That is why on-device safeguarding represents an additional and entirely new category of online protection.
Rather than attempting to regulate access to online services, on-device safeguarding protects children wherever they are on their phone, tablet, gaming platform or laptop. Working in real time, it prevents the viewing, filming and sharing of explicit content before the harm occurs, regardless of which application is being used.
This is prevention, not moderation.
Don't tie protection to age verification
The UK Government's proposal to require technology companies to include on-device safeguarding is an important and welcome development.
Where we disagree is the proposal that these protections should be switched on or off through age verification.
That approach risks undermining public confidence in a technology that has the potential to transform child safety.
Whether justified or not, many people already fear that mandatory age verification represents the thin end of a much larger system of government surveillance. Requiring adults to verify their identity before they can control software on their own devices risks
reinforcing that belief and creating unnecessary resistance to technology whose sole purpose is protecting children.
More importantly, it creates a safeguarding gap.
Children do not only use their own devices. They regularly use those belonging to parents, grandparents, relatives and friends. This is especially relevant for low-income families and communities. If protection has been disabled because the adult has been compelled to verify their age, the child inherits an unprotected device at precisely the moment they need safeguarding.
Protection should follow the child.
Not the identity of the device owner.
Parents should decide
SafeToNet believes responsibility should remain where it has always belonged. With parents.
No government knows a child better than their parents.
No government loves them more.
The role of government is to ensure effective protection is available. The role of parents is to decide when and how that protection should be used.
The law already expects parents to make decisions about seatbelts, child car seats, medicines and countless other safety measures. Digital safety should be no different.
The law should require parents to activate effective protection for children using their devices, while leaving families, not governments, responsible for those decisions.
Independent regulation builds trust
Public confidence will depend upon independent assurance.
Parents should never be asked simply to trust technology companies when they claim their products are private, effective and safe.
Every company developing on-device safeguarding should therefore be independently regulated in every country in which it operates.
Independent assessment should confirm that no personal content ever leaves the device, no hidden backdoors exist, and privacy is fully preserved. It should verify that the AI has been trained using ethically sourced material rather than AI-generated or unlawful datasets, that it operates fairly without bias, that it works accurately across
every application including cameras and livestreams, that it cannot be circumvented, that it does not materially affect battery life or performance, and that it can be deployed across virtually every modern device through manufacturers.
International standards should be agreed so every provider is measured against the same objective benchmarks.
Social media bans alone are not enough
The same principle applies to social media.
The debate has become unnecessarily polarised.
Either ban children from social media or allow unrestricted access.
Reality is more complicated.
Many children, particularly those who are disabled or neurodiverse, rely on online communities for friendship, education and social interaction. Blanket prohibitions fail to recognise that one size rarely fits every child.
Parents should decide when their child is ready to access social media, supported by embedded parental controls and on-device safeguards that continue protecting them after they have logged in.
The objective should not simply be controlling access.
It should be preventing harm.
Protection should be universal
Finally, protection should never depend upon a family's income.
Every child deserves the same level of safety, regardless of where they live or what their parents can afford.
On-device safeguarding should be embedded by manufacturers, independently assured, controlled by parents and free at the point of use.
That is how trust is built.
That is how children are protected.
And that is why on-device safeguarding represents the next step forward in online safety.