Contribution
I agree; my hon. Friend makes a very important point about the slander that happens online, the lack of basis in reality and the lack of ability to address it. If somebody posts something about someone else that is untrue, platforms will not take it down; they will say, “It doesn’t breach our terms and conditions.” Somebody could post that I am actually purple and have pink eyes. I would say, “I don’t want you to say that,” and the platform would say, “But there’s nothing offensive about it.” I would say, “But it’s not me.” The thing is that this is happening in much more offensive ways.
My hon. Friend the Member for Oldham West, Chadderton and Royton (Jim McMahon) made the point that what happens online is then repeated offline. We have even seen deaths when children try to replicate the challenges that they see being set online. With AI-generated material, those challenges often are not real. It is the equivalent of somebody trying to repeat magic tricks and dying as a result, which is quite worrying.
The Online Safety Act is not perfect; it needs to go further. The petitioner has made a really important point. The lack of proper definition around small but non-harmful sites versus small but harmful sites is very unclear, and it is really important that the Act provides some clarity on that.
We do not have enough protections for democracy. The Science, Innovation and Technology Committee, which I am a member of, produced a really important report on misinformation and how it led to the riots two summers ago. Misinformation was used as a rallying cry to create unrest across our country of a sort that we had not seen in a very long time. The response from the social media companies was variable; it went from kind of “meh” to really awful. The platforms say, “We don’t police our content. We’re just a platform.” That is naive in the extreme. Quite frankly, they are happy to make money off us, so they should also know that they have to protect us—their customers—just as any other company does, as my hon. Friend the Member for Oldham West, Chadderton and Royton said.
The radicalisation that is happening online is actually shifting the Overton window; we are seeing a more divided country. There is a fantastic book called “Man Up”—it is very academic, but it shows the rise of misogyny leading to the rise of every other form of extremism and how that links back to the online world. If this was all about Islam, this House would be outraged, but because it starts with misogyny, it goes down with a fizzle, and too often people in this House say, “This is all about free speech.” We know that misogyny is the first step on a ladder of radicalisation that leads people to violence—whether into violence against women or further into antisemitism, anti-Islam, anti-anybody who is not the same colour, or anti-anybody who is perceived not to be English from Norman times.
The algorithms provoke violent and shocking content, but they also shadow-ban really important content, such as information on women’s health. Platforms are happy to shadow-ban terms such as “endometriosis” and “tampon”—and God forbid that a tampon commercial should feature red liquid, rather than blue liquid. That content gets shadow-banned and is regularly taken down and taken out of the algorithms, yet the platforms say they can do nothing about people threatening to rape and harm. That is not true; they can, and they choose not to. The public agree that algorithms must be part of the solution; 78% of British parents want to see action on algorithms. My hon. Friends are right that the Online Safety Act and Ofcom could do that, yet they have not done so—they have yet to create transparency in algorithms, which was the Select Committee’s No. 1 recommendation.
[Sir John Hayes in the Chair]
Finally, I want to talk about a few other areas in which we need to move very quickly: deepfakes and AI nudifying apps. We have already seen an example of how deepfakes are being used in British democracy: a deepfake was made of the hon. Member for Mid Norfolk (George Freeman) saying that he is moving from the Conservatives to Reform. It is a very convincing three-minute video. Facebook still refuses to take it down because it does not breach its terms. This should be a warning to us all about how individuals, state actors and non-state actors can impact our local democracy by creating deepfakes of any one of us that we cannot get taken down.