(2 years, 5 months ago)
Public Bill CommitteesI rise briefly to talk about content reporting. I share the frustrations of the hon. Member for Aberdeen North. The way I read the Bill was that it would allow users and affected persons, rather than “or” affected persons, to report content. I hope the Minister can clarify that that means affected persons who might not be users of a platform. That is really important.
Will the Minister also clarify the use of human judgment in these decisions? Many algorithms are not taking down some content at the moment, so I would be grateful if he clarified that there is a need for platforms to provide a genuine human judgment on whether content is harmful.
I want to raise an additional point about content reporting and complaints procedures. I met with representatives of Mencap yesterday, who raised the issue of the accessibility of the procedures that are in place. I appreciate that the Bill talks about procedures being accessible, but will the Minister give us some comfort about Ofcom looking at the reporting procedures that are in place, to ensure that adults with learning disabilities in particular can access those content reporting and complaints procedures, understand them and easily find them on sites?
That is a specific concern that Mencap raised on behalf of its members. A number of its members will be users of sites such as Facebook, but may find it more difficult than others to access and understand the procedures that are in place. I appreciate that, through the Bill, the Minister is making an attempt to ensure that those procedures are accessible, but I want to make sure they are accessible not just for the general public but for children, who may need jargon-free access to content reporting and complaints procedures, and for people with learning disabilities, who may similarly need jargon-free, easy-to-understand and easy-to-find access to those procedures.
(2 years, 5 months ago)
Public Bill CommitteesI absolutely agree with the points that have been made about the violence against women code of conduct. It is vital, and it would be a really important addition to the Bill. I associate myself with the shadow Minister’s comments, and am happy to stand alongside her.
I want to make a few comments about new clause 20 and some of the issues it raises. The new clause is incredibly important, and we need to take seriously the concerns that have been raised with us by the groups that advocate on behalf of children. They would not raise those concerns if they did not think the Bill was deficient in this area. They do not have spare people and cannot spend lots of time doing unnecessary things, so if they are raising concerns, those are very important things that will make a big difference.
I want to go a little further than what the new clause says and ask the Minister about future-proofing the Bill and ensuring that technologies can be used as they evolve. I am pretty sure that everybody agrees that there should be no space where it is safe to share child sexual exploitation and abuse, whether physical space or online space, private messaging or a more open forum. None of those places should be safe or legal. None should enable that to happen.
My particular thought about future-proofing is about the development of technologies that are able to recognise self-generated pictures, videos, livestreams and so on that have not already been categorised, do not have a hash number and are not easy for the current technologies to find. There are lots of people out there working hard to stamp out these images and videos online, and I have faith that they are developing new technologies that are able to recognise images, videos, messages and oral communications that cannot currently be recognised.
I agree wholeheartedly with the new clause: it is important that a report be produced within six months of the Bill being passed. It would be great if the Minister would commit to thinking about whether Ofcom will be able to require companies to implement new technologies that are developed, as well as the technologies that are currently available. I am not just talking about child sexual abuse images, material or videos; I am also talking about private messaging where grooming is happening. That is a separate thing that needs to be scanned for, but it is incredibly important.
Some of the stories relayed by the shadow Minister relate to conversations and grooming that happened in advance of the self-generated material being created. If there had been a proactive action to scan for grooming behaviour by those companies whose platforms the direct messaging was taking place on, then those young people would potentially have been in a safer place, because it could have been stopped in advance of that self-generated material being created. Surely, that should be the aim. It is good that we can tackle this after the event—it is good that we have something—but tackling it before it happens would be incredibly important.
Online sexual exploitation is a horrific crime, and we all want to see it ended for good. I have concerns about whether new clause 20 is saying we should open up all messaging—where is the consideration of privacy when the scanning is taking place? Forgive me, I do not know much about the technology that is available to scan for that content. I do have concerns that responsible users will have an infringement of privacy, even when doing nothing of concern.
I do not know whether everybody draws the same distinction as me. For me the distinction is that, because it will be happening with proactive technology—technological means will be scanning those messages rather than humans—nobody will see the messages. Software will scan messages, and should there be anything that is illegal—should there be child sexual abuse material—that is what will be flagged and further action taken.
I thank the shadow Minister for her assistance with that intervention, which was incredibly helpful. I do not have concerns that anybody will be able to access that data. The only data that will be accessible is when the proactive technology identifies something that is illegal, so nobody can see any of the messages except for the artificial intelligence. When the AI recognises that something is abuse material, at that point the Bill specifies that it will go to the National Crime Agency if it is in relation to child abuse images.
My concern is that, at the point at which the data is sent to the National Crime Agency, it will be visible to human decision making. I am wondering whether that will stop parents sharing pictures of their babies in the bath? There are instances where people could get caught up in a very innocent situation that is deemed to be something more sinister by AI. However, I will take the advice of the hon. Member for Pontypridd advice and look into the technology.
In terms of the secondary processes that kick in after the AI has scanned the data, I assume it will be up to Ofcom and the provider to discuss what happens then. Once the AI identifies something, does it automatically get sent to the National Crime Agency, or does it go through a process of checking to ensure the AI has correctly identified something? I agree with what the Minister has reiterated on a number of occasions; if it is child sexual abuse material then I have no problem with somebody’s privacy being invaded in order for that to be taken to the relevant authorities and acted on.
I want to make one last point. The wording of new clause 20 is about a report on those proactive technologies. It is about requiring Ofcom to come up with and justify the use of those proactive technologies. To give the hon. Member for Wolverhampton North East some reassurance, it is not saying, “This will definitely happen.” I assume that Ofcom will be able to make the case—I am certain it will be able to—but it will have to justify it in order to be able to require those companies to undertake that use.
My key point is about the future-proofing of this, ensuring that it is not just a one-off, and that, if Ofcom makes a designation about the use of proactive technologies, it is able to make a re-designation or future designation, should new proactive technologies come through, so that we can require those new proactive technologies to be used to identify things that we cannot identify with the current proactive technologies.