ONLINE SAFETY BILL (First sitting)

Sarah Owen Excerpts

Committee stage (re-committed clauses and schedules)
Tuesday 13th December 2022 Public Bill Committees

[Kirsty Blackman (Aberdeen North) (SNP)]

I beg to move amendment 98, in clause 11, page 10, line 17, at end insert

“, and—

(c) mitigate the harm to children caused by habit-forming features of the service by consideration and analysis of how processes (including algorithmic serving of content, the display of other users’ approval of posts and notifications) contribute to development of habit-forming behaviour.”

This amendment requires services to take or use proportionate measures to mitigate the harm to children caused by habit-forming features of a service.

Thank you, Sir Roger, for chairing this recommitted Bill Committee. I will not say that it is nice to be back discussing the Bill again; we had all hoped to have made more progress by now. If you will indulge me for a second, I would like to thank the Clerks, who have been massively helpful in ensuring that this quick turnaround could happen and that we could table the amendments in a sensible form.

Amendment 98 arose from comments and evidence from the Royal College of Psychiatrists highlighting that a number of platforms, and particularly social media platforms such as TikTok and Facebook, generally encourage habit-forming behaviour or have algorithms that encourage it. Such companies are there to make money—that is what companies do—so they want people to linger on their sites and to spend as much time there as possible.

I do not know how many hon. Members have spent time on TikTok, but if they do, and they enjoy some of the cat videos, for instance, the algorithm will know and will show them more videos of cats. They will sit there and think, “Gosh, where did the last half-hour go? I have been watching any number of 20-second videos about cats, because they constantly come up.” Social media sites work by encouraging people to linger on the site and to spend the time dawdling and looking at the advertisements, which make the company additional revenue.

That is good for capitalism and for the company’s ability to make money but the issue, particularly in relation to clause 11, is how that affects children. Children may not have the necessary filters; they may not have the ability that we have to put our phones down—not that we always manage to do so. That ability and decision-making process may not be as refined in children as in adults. Children can be sucked into the platforms by watching videos of cats or of something far more harmful.

[Sarah Owen (Luton North) (Lab)]

The hon. Member makes an excellent point about TikTok, but it also applies to YouTube. The platforms’ addictive nature has to do with the content. A platform does not just show a person a video of a cat, because that will not keep them hooked for half an hour. It has to show them a cat doing something extraordinary, and then a cat doing something even more extraordinary. That is why vulnerable people, especially children, get sucked into a dark hole. They click to see not just the same video but something more exciting, and then something even more exciting. That is the addictive nature of this.

[Kirsty Blackman]

That is absolutely the case. We are talking about cats because I chose them to illustrate the situation, but people may look at content about healthy eating, and that moves on to content that encourages them to be sick. The way the algorithms step it up is insidious; they get more and more extreme, so that the linger time is increased and people do not get bored. It is important that platforms look specifically at their habit-forming features.

--- Later in debate ---

[The Parliamentary Under-Secretary of State for Digital, Culture, Media and Sport (Paul Scully)]

It is a pleasure to serve under your chairmanship, Sir Roger.

Being online can be a hugely positive experience for children and young people, but we recognise the challenge of habit-forming behaviour or designed addiction to some digital services. The Bill as drafted, however, would already deliver the intent of the amendment from the hon. Member for Aberdeen North. If service providers identify in their risk assessment that habit-forming or addictive-behaviour risks cause significant harm to an appreciable number of children on a service, the Bill will require them to put in place measures to mitigate and manage that risk under clause 11(2)(a).

To meet the child safety risk assessment duties under clause 10, services must assess the risk of harm to children from the different ways in which the service is used; the impact of such use; the level of risk of harm to children; how the design and operation of the service may increase the risks identified; and the functionalities that facilitate the presence or dissemination of content of harm to children. The definition of “functionality” at clause 200 already includes an expression of a view on content, such as applying a “like” or “dislike” button, as at subsection (2)(f)(i).

[Sarah Owen]

I thank the Minister for giving way so early on. He mentioned an “appreciable number”. Will he clarify what that is? Is it one, 10, 100 or 1,000?

[Paul Scully]

I do not think that a single number can be put on that, because it depends on the platform and the type of viewing. It is not easy to put a single number on that. An “appreciable number” is basically as identified by Ofcom, which will be the arbiter of all this. It comes back to what the hon. Member for Aberdeen North said about the direction that we, as she rightly said, want to give Ofcom. Ofcom has a range of powers already to help it assess whether companies are fulfilling their duties, including the power to require information about the operation of their algorithms. I would set the direction that the hon. Lady is looking for, to ensure that Ofcom uses those powers to the fullest and can look at the algorithms. We should bear in mind that social media platforms face criminal liability if they do not supply the information required by Ofcom to look under the bonnet.

--- Later in debate ---

[Kirsty Blackman]

That is exactly why users should be able to block private messaging in general. Someone on Twitter can say, “I’m not going to receive a direct message from anybody I don’t follow.” Twitter users have the opportunity to do that, but there is not necessarily that opportunity on all platforms. We are asking for those things to be included, so that the provider can say, “You’re using private messaging inappropriately. Therefore, we are blocking all your access to private messaging,” or, “You are being harmed as a result of accessing private messaging. Therefore, we are blocking your access to any private messaging. You can still see pictures on Instagram, but you can no longer receive any private messages, because we are blocking your access to that part of the site.” That is very different from blocking a user’s access to certain kinds of content, for example. I agree that that should happen, but it is about the functionalities and stopping access to some of them.

We are not asking Ofcom to mandate that platforms take this measure; they could still take the slightly more nuclear option of banning somebody entirely from their service. However, if this option is included, we could say, “Your service is doing pretty well, but we know there is an issue with private messaging. Could you please take action to ensure that those people who are using private messaging to harm children no longer have access to private messaging and are no longer able to use the part of the service that enables them to do these things?” Somebody might be doing a great job of making games in Roblox, but they may be saying inappropriate things. It may be proportionate to block that person entirely, but it may be more proportionate to block their access to voice chat, so that they can no longer say those things, or direct message or contact anybody. It is about proportionality and recognising that the service is not necessarily inherently harmful but that specific parts of it could be.

[Sarah Owen]

The hon. Member is making fantastic, salient points. The damage with private messaging is around phishing, as well as seeing a really harmful message and not being able to unsee it. Would she agree that it is about protecting the victim, not putting the onus on the victim to disengage from such conversations?

[Kirsty Blackman]

I completely agree. The hon. Member put that much better than I could. I was trying to formulate that point in my head, but had not quite got there, so I appreciate her intervention. She is right: we should not put the onus on a victim to deal with a situation. Once they have seen a message from someone, they can absolutely block that person, but that person could create another account and send them messages again. People could be able to choose, and to say, “No, I don’t want anyone to be able to send me private messages,” or “I don’t want any private messages from anyone I don’t know.” We could put in those safeguards.

I am talking about adding another layer to the clause, so that companies would not necessarily have to demonstrate that it was proportionate to ban a person from using their service, as that may be too high a bar—a concern I will come to later. They could, however, demonstrate that it was proportionate to ban a person from using private messaging services, or from accessing livestreaming features. There has been a massive increase in self-generated child sexual abuse images, and huge amount has come from livestreaming. There are massive risks with livestreaming features on services.

Livestreaming is not always bad. Someone could livestream themselves showing how to make pancakes. There is no issue with that—that is grand—but livestreaming is being used by bad actors to manipulate children into sharing videos of themselves, and once they are on the internet, they are there forever. It cannot be undone. If we were able to ban vulnerable users—my preferred option would be all children—from accessing livestreaming services, they would be much safer.