The internet’s oldest problem is also its youngest: Child safety

For almost 25 years, there has been a major effort to address the issue of children’s safety on the internet. But a major obstacle has continuously impeded this effort: effectively identifying who is an adult and who is a child. 

The modern era of internet policy-making may have begun on July 3, 1995, when America woke up to a spectacular cover of Time magazine detailing a new threat to our children. The cover showed a close-up of young child with a shocked look on his face as he used something called “the Internet” to watch what the magazine cover called, in all caps. “CYBERPORN.”

Until then, the internet was barely known to most Americans. Three subscription-based online networks — AOL, CompuServe and Prodigy — had been around for a few years, but the internet and the worldwide web were used by few Americans and even fewer non-Americans. A text-only worldwide web (what most people think of as the internet) had only been around for about four to five years, and a Web with photos and for just one or two years.  

The 1995 national cyberporn scare led politicians, regulators, educators and law enforcement officials to announce that they would put a stop to children being fed such materials. Many in the computer and online industries responded, including myself, then a computer industry executive responsible for internet policy and regulation. I helped lead a computer industry effort to stop cyberporn from getting to children, years before Google or Facebook. 

It was clear then that the internet would open up a new world of personal anonymity. It was also clear that there was a substantial market for pornography in the U.S., which was generally legal for adults. Finding an effective method to prevent children from seeing cyberporn without blocking it for all adults would be challenging, unlike preventing children from buying pornographic magazines over the counter, where a clerk could easily refuse the sale.  

Thirty years later, the issues involved in protecting children on the Internet have grown far more diverse, complex and international. But much of the challenge of finding effective methods to prevent internet content harmful to children from being accessible to children, while allowing that same content to be accessible to adults, remains. 

One scheme that emerged from early industry cyberporn discussions illustrates the problems. Someone had the idea to require every American adult who wanted to use the Internet to register at their local Post Office and be issued an encrypted cyber-key. The adult would need to use this key to log onto the Internet, which would allow access to any website, pornographic or not. This restriction was supposed to have meant that children could have unsupervised internet access.  

The enormous privacy, free speech, administrative and economic burdens of this plan were so outrageous that it was dropped. The industry effort moved toward technical and market-based approaches to protecting children from cyberporn, such as a voluntary website ratings system patterned after the motion picture industry’s self-rating system for movies.   

Today, we see the same privacy, free speech and technical problems the industry struggled with when the issue was mainly cyberporn, but on a global scale. Among the many children’s Internet issues today are children’s privacy, sexual exploitation, online enticement, sex trafficking, online obscenity, sextortion and the sexual abuse of children. And while the policy mosaic of regulations around the world surrounding these issues can be quite complex, most still must deal with the techniques and implications of distinguishing an adult from a child on the Internet. 

Essentially no technique for identifying and protecting a child on the Internet can be 100 percent effective, any more than real-world techniques can be 100 percent effective. When pornographic magazines were commonly sold over store counters, older-looking teenagers undoubtedly were still able to purchase them. And it has never been difficult for an unscrupulous adult to purchase pornography and pass it on to teens.

Similarly, even checking IDs at a bar does not fully prevent an older child from buying alcohol using something that looks like a valid government ID. So various measures short of the issuance of universal internet IDs have reemerged, despite being dismissed decades ago. None are perfect and, to some degree, all will have some effect on adults’ speech and privacy.  

Consequently, the taxonomy of identifying children on the internet remains complex, imperfect and subject to constant debate, ranging from self-identifying to more complex “trusted third party" verification services that can require photo IDs, videos and more.

Understandably, many who care about children’s internet concerns believe that because large internet businesses make billions of dollars, they should be able to find a way to distinguish between a child and an adult online regardless of the difficulty or cost. Conversely, many civil libertarians, industry interests and technical experts, who have no less concern for children, conclude that such verification measures would fundamentally undermine both the free speech and privacy rights that are essential to a democracy and an open society.  

Some part of this debate is about balancing values (that is, protecting children and preserving privacy and free speech). Some part is about what can and cannot technologically be done. Another part is about profits and global competition. Yet another part is about raw politics and posturing. And some part is about essential intergovernmental coordination.

As policymakers face these issues, it’s important that they strike a balance that reflects their values and assessments as much as their political calculations. 

Roger Cochetti has served as a senior executive with COMSAT, IBM, VeriSign and CompTIA. A former U.S. government official, he has helped found a number of nonprofits in the tech sector and is the author of textbooks on the history of satellite communications.