Ms Niamh Sweeney:

We are investing heavily in artificial intelligence. There are limits to what it can achieve. We see great potential there, especially for tackling illegal content online. It is a bit tricky in some of the examples the Deputy gave, such as child exploitation imagery or images that were shared non-consensually. For that we use photograph matching technology which can prevent their re-upload. We would like to spread the message today that if somebody is threatening or blackmailing a person with sharing a photo, the person should contact us. If we bank that image, we can prevent it from being uploaded. We can do that repeatedly.

It is harder for us to identify those images using artificial intelligence. We are working on that but the difference between an image of somebody sunbathing and pornography is difficult to make. We are investing a lot of time in it. In respect of text, there has been heavy investment in artificial intelligence to combat terrorism. One of the challenges we face is trying to surface material that promotes terrorism as opposed to highlighting the negative effects of terrorism or news content about terrorism. We would surface many false positives if we were to filter for certain words or phrases. In the past ten days we have released some figures on the content that we have managed to remove before it is ever flagged because we have invested in this. There are, however, limits on what can be achieved, for example, a news item where somebody condemns rape versus somebody who has threatened it. All that material would surface at the same time. That is why we are heavily dependent on user reports and we strongly encourage people to take action where they can.