Top Breakthroughs in Internet Safety

The scale of online threats is growing and cybercriminals are becoming more sophisticated, with the Internet Watch Foundation (IWF) finding over 252,000 URLs containing child sexual abuse material (CSAM) in recent years. As stated by the World Economic Forum, platforms are being weaponized to livestream terrorist attacks, cyberbullying rates are through the roof, and in some Western countries, around 67% of people aged 13 to 24 have encountered potentially harmful content online. The good news is that technologies like machine learning are doing plenty to give Internet users peace of mind. Below are just a few new breakthroughs in this sector.

Client-Side Scanning

Client-side scanning (CSS) enables systems to scan pictures, text, or video in a message and match it to a database of illegal or inappropriate content before the message reaches the intended recipient on an encrypted channel. This technology leaves all encrypted parts of a system intact, thus maintaining users’ privacy.  It is noteworthy that some see potential problems with CSS, including the potential expansion of its scope. Although CSS systems are designed to scan for CSAM, some experts believe that criminal entities could undermine these systems and use the technology to detect private personal or business data or to insert false positives against political enemies.

Artificial Intelligence and Natural Language Processing (NLP) Models

AI systems can be used to automate content moderation processes and detect harmful content via NLP models. Currently, the main challenge for this technology is that different platforms have different sets of users, content types, and vocabulary. Some models are now capable of “zero-shot” and “few-shot” learning, which allows them to perform well even with a small amount of data. The difference between these two learning types is that while few-shot learning enables models to learn from a very small number of labeled examples per class, zero-shot learning allows models to recognize objects they have never encountered before during training. The hope is that these advances will enable developers to create scalable AI that can be personalized to each platform.

New Blocking Apps

A recent report shows that 73% of teens aged 13 to 17 have watched pornography online, and more than half reported encountering this type of material before they were 13. Parents naturally wish to safeguard their kids against this type of content and they now have access to a wide variety of apps. This means that they can easily find a web blocker for Chrome, Microsoft Edge, or Firefox, which they can install on desktop and mobile devices. Today, parents can not only block specific sites and content but also limit access to games to specific times. Kids can access the Internet for study purposes even though their use of gaming sites like Fortnite can be temporarily restricted.

Image Recognition

All images uploaded to the Internet and shared can now be recognized thanks to a system that uses an algorithm to assign a unique hash value to each of them. Any copies of an image will have the same hash value, which makes it easy for security systems to identify unsavory or illegal content. The latest versions of this technology allow for the detection of edited images. That means that even if an image is altered, the system can still detect that it is the same as the original. 

Biometrics to Identify Age and Identity

Biometrics—which relies on voice, iris, and face scans to identify people—is an excellent way to verify factors such as age and identity. In the future, users will choose which identifying feature they wish to provide to access a site, platform, or app. For instance, some might prefer iris scans while others prefer to record their voice. Dating sites like Tinder have already adopted this technology to boost trust and safety. Age verification, meanwhile, can ensure that children are not exposed to adult-only content. Facial analysis technology investigates aspects such as gray hair, sun spots, and facial lines to estimate the user’s age, thus helping to keep minors safe.

Breakthroughs such as client-side scanning, advanced image recognition, and identity verification are all groundbreaking technologies in curbing cybercriminal activities and harmful content. Although concerns about privacy and scalability remain, these innovations are helping to create a safer digital environment for users. This is particularly true for children and teens, who spend a large portion of their day online.