Sharon Keogan (Independent) | Oireachtas source

Later today, the European Commission is set to publish its draft law on so-called chat control, which will oblige providers of digital correspondence platforms to search all private chat messages and emails automatically for suspicious content. This searching will be done by way of algorithm artificial intelligence, AI, and carried out in a blanket manner. Everyone's messages will be scanned all the time, which will result in the end of secure online communications within the EU. If an algorithm considers a message suspicious, its content and metadata are disclosed automatically and without human verification to a private US-based organisation. The reported users are not notified. While this may seem to be a case that if a person has nothing to hide, he or she has nothing to fear, there are many grave concerns with this move towards what has been seen as Chinese Government-style surveillance.

The aim of this legislation is to combat the sharing of child sexual exploitation material. The experience of national and international agencies that combat this material online, however, shows that mass surveillance is the wrong approach to fighting child abuse material and sexual exploitation. This is because abusers do not share their material via commercial email messaging or chat services but organise themselves through self-run secret forums on the darknet, which would not be caught by these controlled algorithms. Abusers also typically upload images and videos as encrypted archives and share only the links and passwords, which these chat-controlled algorithms would not recognise.

The implications for the privacy of 500 million people are stark. We in this House passed a law last year to criminalise the sharing of intimate photographs without consent yet if an algorithm now classifies the content of a message as suspicious, an individual's private or intimate photographs may be viewed by staff and contractors of private international corporations and police authorities.

According to Swiss federal police authorities, 86% of all machine-generated reports turn out to be without merit. We need a human task force to tackle this problem. Of course, 72% of EU citizens oppose the measure, as per a YouGov poll of more than 10,000 people last year, with only 18% supporting the plan. Again, this is something that is coming down the line. We ought to have debate on what we want our standard of online privacy to be for our citizens.