Online Safety Bill (Sixth sitting)

Kirsty Blackman Excerpts

Committee stage
Tuesday 7th June 2022 Public Bill Committees

[The Chair]

Welcome back. I have a few announcements. I have been reassured that we will have no transmission problems this afternoon, and apparently the audio of this morning’s sitting is available if Members want to listen to it. I have no objections to Members taking their jackets off, because it is rather warm this afternoon. We are expecting a Division in the main Chamber at about 4 o’clock, so we will suspend for 15 minutes if that happens.

[Kirsty Blackman (Aberdeen North) (SNP)]

I am sorry, Ms Rees, but I am afraid that I cannot hear you very well.

[The Chair]

I will shout a bit in that case.

Clause 8

Illegal content risk assessment duties

Amendment proposed (this day): 10, in clause 8, page 6, line 33, at end insert—

“(4A) A duty to publish the illegal content risk assessment and proactively supply this to OFCOM.”—(Alex Davies-Jones.)

This amendment creates a duty to publish an illegal content risk assessment and supply it to Ofcom.

Question again proposed, That the amendment be made.

--- Later in debate ---

[Chris Philp]

I thank my right hon. Friend for raising that. The risk assessments and, indeed, the duties arising under this Bill all apply to systems and processes—setting up systems and processes that are designed to protect people and to prevent harmful and illegal content from being encountered. We cannot specify in legislation every type of harmful content that might be encountered. This is about systems and processes. We heard the Chairman of the Joint Committee on the draft Online Safety Bill, our hon. Friend the Member for Folkestone and Hythe (Damian Collins), confirm to the House on Second Reading his belief—his accurate belief—that the Bill takes a systems-and-processes approach. We heard some witnesses saying that as well. The whole point of this Bill is that it is tech-agnostic—to future-proof it, as hon. Members mentioned this morning—and it is based on systems and processes. That is the core architecture of the legislation that we are debating.

Amendments 25 and 26 seek to ensure that user-to-user services assess and mitigate the risk of illegal content being produced via functions of the service. That is covered, as it should be—the Opposition are quite right to raise the point—by the illegal content risk assessment and safety duties in clauses 8 and 9. Specifically, clause 8(5)(d), on page 7 of the Bill—goodness, we are only on page 7 and we have been going for over half a day already—requires services to risk-assess functionalities of their service being used to facilitate the presence of illegal content. I stress the word “presence” in clause 8(5)(d). Where illegal content is produced by a functionality of the service—for example, by being livestreamed—that content will be present on the service and companies must mitigate that risk. The objective that the Opposition are seeking to achieve, and with which we completely agree with, is covered in clause 8(5)(d) by the word “presence”. If the content is present, it is covered by that section.

[Kirsty Blackman]

Specifically on that, I understand the point the hon. Gentleman is making and appreciate his clarification. However, on something such as Snapchat, if somebody takes a photo, it is sent to somebody else, then disappears immediately, because that is what Snapchat does—the photo is no longer present. It has been produced and created there, but it is not present on the platform. Can the Minister consider whether the Bill adequately covers all the instances he hopes are covered?

[Chris Philp]

The hon. Lady raises an interesting point about time. However, the clause 8(5)(d) uses the wording,

“the level of risk of functionalities of the service facilitating the presence or dissemination of illegal content”

and so on. That presence can happen at any time, even fleetingly, as with Snapchat. Even when the image self-deletes after a certain period—so I am told, I have not actually used Snapchat—the presence has occurred. Therefore, that would be covered by clause 8(5)(d).

[Chris Philp]

The question of proof is a separate one, and that would apply however we drafted the clause. The point is that the clause provides that any presence of a prohibited image would fall foul of the clause. There are also duties on the platforms to take reasonable steps. In the case of matters such as child sexual exploitation and abuse images, there are extra-onerous duties that we have discussed before, for obvious and quite correct reasons.

[Kirsty Blackman]

Will the Minister stress again that in this clause specifically he is talking about facilitating any presence? That is the wording that he has just used. Can he clarify exactly what he means? If the Minister were to do so, it would be an important point for the Bill as it proceeds.

[The Chair]

Order. Minister, before you continue, before the Committee rose earlier today, there was a conversation about clause 9 being in, and then I was told it was out. This is like the hokey cokey; it is back in again, just to confuse matters further. I was confused enough, so that point needs to be clarified.

--- Later in debate ---

[Alex Davies-Jones]

I am grateful for the opportunity to speak to amendments to clause 9 and to clauses 23 and 24, which I did not speak on earlier. I am also very grateful that we are being broadcast live to the world and welcome that transparency for all who might be listening.

On clause 9, it is right that the user-to-user services will be required to have specific duties and to take appropriate measures to mitigate and manage the risk of harm to individuals and their likelihood of encountering priority illegal content. Again, however, the Bill does not go far enough, which is why we are seeking to make these important amendments. On amendment 18, it is important to stress that the current scope of the Bill does not capture the range of ways in which child abusers use social networks to organise abuse, including to form offender networks. They post digital breadcrumbs that signpost to illegal content on third-party messaging apps and the dark web, and they share child abuse videos that are carefully edited to fall within content moderation guidelines. This range of techniques, known as child abuse breadcrumbing, is a significant enabler of online child abuse.

Our amendment would give the regulator powers to tackle breadcrumbing and ensure a proactive upstream response. The amendment would ensure that tens of millions of interactions with accounts that actively enable the discovery and sharing of child abuse material will be brought into regulatory scope. It will not leave that as ambiguous. The amendment will also ensure that companies must tackle child abuse at the earliest possible stage. As it stands, the Bill would reinforce companies’ current focus only on material that explicitly reaches the criminal threshold. Because companies do not focus their approach on other child abuse material, abusers can exploit this knowledge to post carefully edited child abuse images and content that enables them to connect and form networks with other abusers. Offenders understand and can anticipate that breadcrumbing material will not be proactively identified or removed by the host site, so they are able to organise and link to child abuse in plain sight.

We all know that child abuse breadcrumbing takes many forms, but techniques include tribute sites where users create social media profiles using misappropriated identities of known child abuse survivors. These are used by offenders to connect with likeminded perpetrators to exchange contact information, form offender networks and signpost child abuse material elsewhere online. In the first quarter of 2021, there were 6 million interactions with such accounts.

Abusers may also use Facebook groups to build offender groups and signpost to child abuse hosted on third-party sites. Those groups are thinly veiled in their intentions; for example, as we heard in evidence sessions, groups are formed for those with an interest in children celebrating their 8th, 9th or 10th birthdays. Several groups with over 50,000 members remained alive despite being reported to Meta, and algorithmic recommendations quickly suggested additional groups for those members to join.

Lastly, abusers can signpost to content on third-party sites. Abusers are increasingly using novel forms of technology to signpost to online child abuse, including QR codes, immersive technologies such as the metaverse, and links to child abuse hosted on the blockchain. Given the highly agile nature of the child abuse threat and the demonstrable ability of sophisticated offenders to exploit new forms of technology, this amendment will ensure that the legislation is effectively futureproofed. Technological change makes it increasingly important that the ability of child abusers to connect and form offender networks can be disrupted at the earliest possible stage.

Turning to amendment 21, we know that child abuse is rarely siloed on a single platform or app. Well-established grooming pathways see abusers exploit the design features of social networks to contact children before they move communication across to other platforms, including livestreaming sites, as we have already heard, and encrypted messaging services. Offenders manipulate features such as Facebook’s algorithmic friend suggestions to make initial contact with a large number of children. They can then use direct messages to groom them and coerce children into sending sexual images via WhatsApp. Similarly, as we heard earlier, abusers can groom children through playing videogames and then bringing them on to another ancillary platform, such as Discord.

The National Society for the Prevention of Cruelty to Children has shared details of an individual whose name has been changed, and whose case particularly highlights the problems that children are facing in the online space. Ben was 14 when he was tricked on Facebook into thinking he was speaking to a female friend of a friend, who turned out to be a man. Using threats and blackmail, he coerced Ben into sending abuse images and performing sex acts live on Skype. Those images and videos were shared with five other men, who then bombarded Ben with further demands. His mum, Rachel, said:

“The abuse Ben suffered had a devastating impact on our family. It lasted two long years, leaving him suicidal.

It should not be so easy for an adult to meet and groom a child on one site then trick them into livestreaming their own abuse on another app, before sharing the images with like-minded criminals at the click of a button.

Social media sites should have to work together to stop this abuse happening in the first place, so other children do not have to go through what Ben did.”

The current drafting of the Bill does not place sufficiently clear obligations on platforms to co-operate on the cross-platform nature of child abuse. Amendment 21 would require companies to take reasonable and proportionate steps to share threat assessments, develop proportionate mechanisms to share offender intelligence, and create a rapid response arrangement to ensure that platforms develop a coherent, systemic approach to new and emerging threats. Although the industry has developed a systemic response to the removal of known child abuse images, these are largely ad hoc arrangements that share information on highly agile risk profiles. The cross-platform nature of grooming and the interplay of harms across multiple services need to be taken into account. If it is not addressed explicitly in the Bill, we are concerned that companies may be able to cite competition concerns to avoid taking action.

[Kirsty Blackman]

On the topic of child abuse images, the hon. Member spoke earlier about livestreaming and those images not being captured. I assume that she would make the same point in relation to this issue: these live images may not be captured by AI scraping for them, so it is really important that they are included in the Bill in some way as well.

[Alex Davies-Jones]

I completely agree with the hon. Member, and appreciate her intervention. It is fundamental for this point to be captured in the Bill because, as we are seeing, this is happening more and more. More and more victims are coming forward who have been subject to livestreaming that is not picked up by the technology available, and is then recorded and posted elsewhere on smaller platforms.

Legal advice suggests that cross-platform co-operation is likely to be significantly impeded by the negative interplay with competition law unless there is a clear statutory basis for enabling or requiring collaboration. Companies may legitimately have different risk and compliance appetites, or may simply choose to hide behind competition law to avoid taking a more robust form of action.

New and emerging technologies are likely to produce an intensification of cross-platform risks in the years ahead, and we are particularly concerned about the child abuse impacts in immersive virtual reality and alternative-reality environments, including the metaverse. A number of high-risk immersive products are already designed to be platform-agnostic, meaning that in-product communication takes place between users across multiple products and environments. There is a growing expectation that these environments will be built along such lines, with an incentive for companies to design products in this way in the hope of blunting the ability of Governments to pursue user safety objectives.

Separately, regulatory measures that are being developed in the EU, but are highly likely to impact service users in the UK, could result in significant unintended safety consequences. Although the interoperability provisions in the Digital Markets Act are strongly beneficial when viewed through a competition lens—they will allow the competition and communication of multiple platforms—they could, without appropriate safety mitigations, provide new means for abusers to contact children across multiple platforms, significantly increase the overall profile of cross-platform risk, and actively frustrate a broad number of current online safety responses. Amendment 21 will provide corresponding safety requirements that can mitigate the otherwise significant potential for unintended consequences.

The Minister referred to clauses 23 and 24 in relation to amendments 30 and 31. We think a similar consideration should apply for search services as well as for user-to-user services. We implore that the amendments be made, in order to prevent those harms from occurring.

--- Later in debate ---

[Chris Philp]

The Government support the spirit of amendments 17 and 28, which seek to achieve critical objectives, but the Bill as drafted delivers those objectives. In relation to amendment 17 and cross-platform risk, clause 8 already sets out harms and risks—including CSEA risks—that arise by means of the service. That means through the service to other services, as well as on the service itself, so that is covered.

Amendment 28 calls for the risk assessments expressly to cover illegal child sexual exploitation content, but clause 8 already requires that to happen. Clause 8(5) states that the risk assessment must cover the

“risk of individuals who are users of the service encountering…each kind of priority illegal content”.

If we follow through the definition of priority illegal content, we find all those CSEA offences listed in schedule 6. The objective of amendment 28 is categorically delivered by clause 8(5)(b), referencing onwards to schedule 6.

[Kirsty Blackman]

The amendment specifically mentions the level and rates of those images. I did not quite manage to follow through all the things that the Minister just spoke about, but does the clause specifically talk about the level of those things, rather than individual incidents, the possibility of incidents or some sort of threshold for incidents, as in some parts of the Bill?

[Chris Philp]

The risk assessments that clause 8 requires have to be suitable and sufficient; they cannot be perfunctory and inadequate in nature. I would say that suitable and sufficient means they must go into the kind of detail that the hon. Lady requests. More details, most of which relate to timing, are set out in schedule 3. Ofcom will be making sure that these risk assessments are not perfunctory.

Importantly, in relation to CSEA reporting, clause 59, which we will come to, places a mandatory requirement on in-scope companies to report to the National Crime Agency all CSEA content that they detect on their platforms, if it has not already been reported. Not only is that covered by the risk assessments, but there is a criminal reporting requirement here. Although the objectives of amendments 17 and 28 are very important, I submit to the Committee that the Bill delivers the intention behind them already, so I ask the shadow Minister to withdraw them.

Question put, That the amendment be made.

--- Later in debate ---

[Chris Philp]

My right hon. Friend, as always, makes a very good point. The codes of practice will be important, particularly to enable Ofcom to levy fines where appropriate and then successfully defend them. This is an area that may get litigated. I hope that, should lawyers litigating these cases look at our transcripts in the future, they will see how strongly those on both sides of the House feel about this point. I know that Ofcom will ensure that the codes of practice are properly drafted. We touched this morning on the point about timing; we will follow up with Ofcom to make sure that the promise it made us during the evidence session about the road map is followed through and that those get published in good time.

On the point about the Joint Committee, I commend my right hon. Friend for her persistence—[Interruption.] Her tenacity—that is the right word. I commend her for her tenacity in raising that point. I mentioned it to the Secretary of State when I saw her at lunchtime, so the point that my right hon. Friend made this morning has been conveyed to the highest levels in the Department.

I must move on to the final two amendments, 11 and 13, which relate to transparency. Again, we had a debate about transparency earlier, when I made the point about the duties in clause 64, which I think cover the issue. Obviously, we are not debating clause 64 now but it is relevant because it requires Ofcom—it is not an option but an obligation; Ofcom must do so—to require providers to produce a transparency report every year. Ofcom can say what is supposed to be in the report, but the relevant schedule lists all the things that can be in it, and covers absolutely everything that the shadow Minister and the hon. Member for Worsley and Eccles South want to see in there.

That requirement to publish transparently and publicly is in the Bill, but it is to be found in clause 64. While I agree with the Opposition’s objectives on this point, I respectfully say that those objectives are delivered by the Bill as drafted, so I politely and gently request that the amendments be withdrawn.

[Kirsty Blackman]

I have a couple of comments, particularly about amendments 15 and 16, which the Minister has just spoken about at some length. I do not agree with the Government’s assessment that the governance subsection is adequate. It states that the risk assessment must take into account

“how the design and operation of the service (including the business model, governance, use of proactive technology…may reduce or increase the risks identified.”

It is actually an assessment of whether the governance structure has an impact on the risk assessment. It has no impact whatever on the level at which the risk assessment is approved or not approved; it is about the risks that the governance structure poses to children or adults, depending on which section of the Bill we are looking at.

The Minister should consider what is being asked in the amendment, which is about the decision-making level at which the risk assessments are approved. I know the Minister has spoken already, but some clarification would be welcome. Does he expect a junior tech support member of staff, or a junior member of the legal team, to write the risk assessment and then put it in a cupboard? Or perhaps they approve it themselves and then nothing happens with it until Ofcom asks for it. Does he think that Ofcom would look unfavourably on behaviour like that? If he was very clear with us about that, it might put our minds at rest. Does he think that someone in a managerial position or a board member, or the board itself, should take decisions, rather than a very junior member of staff? There is a big spread of people who could be taking decisions. If he could give us an indication of what Ofcom might look favourably on, it would be incredibly helpful for our deliberations.

[Chris Philp]

I am anxious about time, but I will respond to that point because it is an important one. The hon. Lady is right to say that clause 10(6)(h) looks to identify the risks associated with governance. That is correct —it is a risk assessment. However in clause 11(2)(a), there is a duty to mitigate those risks, having identified what the risks are. If, as she hypothesised, a very junior person was looking at these matters from a governance point of view, that would be identified as a risk. If it was not, Ofcom would find that that was not sufficient or suitable. That would breach clause 10(2), and the service would then be required to mitigate. If it did not mitigate the risks by having a more senior person taking the decision, Ofcom would take enforcement action for its failure under clause 11(2)(a).

For the record, should Ofcom or lawyers consult the transcript to ascertain Parliament’s intention in the course of future litigation, it is absolutely the Government’s view, as I think it is the hon. Lady’s, that a suitable level of decision making for a children’s risk assessment would be a very senior level. The official Opposition clearly think that, because they have put it in their amendment. I am happy to confirm that, as a Minister, I think that. Obviously the hon. Lady, who speaks for the SNP, does too. If the transcripts of the Committee’s proceedings are examined in the future to ascertain Parliament’s intention, Parliament’s intention will be very clear.

--- Later in debate ---

[Barbara Keeley]

May I say—this might be a point of order—how my constituency name is pronounced? I get a million different versions, but it is Worsley, as in “worse”. It is an unfortunate name for a great place.

I will speak to all the amendments in the group together, because they relate to how levels of risk are assessed in relation to certain characteristics. The amendments are important because small changes to the descriptions of risk assessment will help to close a significant gap in protection.

Clauses 10 and 12 introduce a duty on regulated companies to assess harms to adults and children who might have an innate vulnerability arising from being a member of a particular group or having a certain characteristic. However, Ofcom is not required to assess harms to people other than children who have that increased innate vulnerability. Amendment 71 would require Ofcom to assess risks of harm particularly affecting people with certain characteristics or membership of a group or groups as part of its risk register. That would reduce the regulatory burden if companies had Ofcom’s risk assessment to base their work on.

Getting this right is important. The risk management regime introduced by the Bill should not assume that all people are at the same risk of harm—they are clearly not. Differences in innate vulnerability increase the incidence and impact of harm, such as by increasing the likelihood of encountering content or of that content being harmful, or heightening the impact of the harm.

It is right that the Bill emphasises the vulnerability of children, but there are other, larger groups with innate vulnerability to online harm. As we know, that often reflects structural inequalities in society.

For example, women will be harmed in circumstances where men might not be, and they could suffer some harms that have a more serious impact than they might for men. A similar point can be made for people with other characteristics. Vulnerability is then compounded by intersectional issues—people might belong to more than one high-risk group—and I will come to that in a moment.

The initial Ofcom risk assessment introduced by clause 83 is not required to consider the heightened risks to different groups of people, but companies are required to assess that risk in their own risk assessments for children and adults. They need to be given direction by an assessment by Ofcom, which amendment 71 would require.

Amendments 72 to 75 address the lack of recognition in these clauses of intersectionality issues. They are small amendments in the spirit of the Bill’s risk management regime. As drafted, the Bill refers to a singular “group” or “characteristic” for companies to assess for risk. However, some people are subject to increased risks of harm arising from being members of more than one group. Companies’ risk assessments for children and adults should reflect intersectionality, and not just characteristics taken individually. Including the plural of “group” and “characteristic” in appropriate places would achieve that.

[Kirsty Blackman]

I will first speak to our amendment 85, which, like the Labour amendment, seeks to ensure that the Bill is crystal clear in addressing intersectionality. We need only consider the abuse faced by groups of MPs to understand why that is necessary. Female MPs are attacked online much more regularly than male MPs, and the situation is compounded if they have another minority characteristic. For instance, if they are gay or black, they are even more likely to be attacked. In fact, the MP who is most likely to be attacked is black and female. There are very few black female MPs, so it is not because of sheer numbers that they are at such increased risk of attack. Those with a minority characteristic are at higher risk of online harm, but the risk facing those with more than one minority characteristic is substantially higher, and that is what the amendment seeks to address.

I have spoken specifically about people being attacked on Twitter, Facebook and other social media platforms, but people in certain groups face an additional significant risk. If a young gay woman does not have a community around her, or if a young trans person does not know anybody else who is trans, they are much more likely to use the internet to reach out, to try to find people who are like them, to try to understand. If they are not accepted by their family, school or workplace, they are much more likely to go online to find a community and support—to find what is out there in terms of assistance—but using the internet as a vulnerable, at-risk person puts them at much more significant risk. This goes back to my earlier arguments about people requiring anonymity to protect themselves when using the internet to find their way through a difficult situation in which they have no role models.

It should not be difficult for the Government to accept this amendment. They should consider it carefully and understand that all of us on the Opposition Benches are making a really reasonable proposal. This is not about saying that someone with only one protected characteristic is not at risk; it is about recognising the intersectionality of risk and the fact that the risk faced by those who fit into more than one minority group is much higher than that faced by those who fit into just one. This is not about taking anything away from the Bill; it is about strengthening it and ensuring that organisations listen.

We have heard that a number of companies are not providing the protection that Members across the House would like them to provide against child sexual abuse. The governing structures, risk assessments, rules and moderation at those sites are better at ensuring that the providers make money than they are at providing protection. When regulated providers assess risk, it is not too much to ask them to consider not just people with one protected characteristic but those with multiple protected characteristics.

As MPs, we work on that basis every day. Across Scotland and the UK, we support our constituents as individuals and as groups. When protected characteristics intersect, we find ourselves standing in Parliament, shouting strongly on behalf of those affected and giving them our strongest backing, because we know that that intersection of harms is the point at which people are most vulnerable, in both the real and the online world. Will the Minister consider widening the provision so that it takes intersectionality into account and not only covers people with one protected characteristic but includes an over and above duty? I genuinely do not think it is too much for us to ask providers, particularly the biggest ones, to make this change.

[Chris Philp]

Once again, the Government recognise the intent behind these amendments and support the concept that people with multiple intersecting characteristics, or those who are members of multiple groups, may experience—or probably do experience—elevated levels of harm and abuse online compared with others. We completely understand and accept that point, as clearly laid out by the hon. Member for Aberdeen North.

There is a technical legal reason why the use of the singular characteristic and group singular is adopted here. Section 6(c) of the Interpretation Act 1978 sets out how words in Bills and Acts are interpreted, namely that such words in the singular also cover the plural. That means that references in the singular, such as

“individuals with a certain characteristic”

in clause 10(6)(d), also cover characteristics in the plural. A reference to the singular implies a reference to the plural.

Will those compounded risks, where they exist, be taken into account? The answer is yes, because the assessments must assess the risk in front of them. Where there is evidence that multiple protected characteristics or the membership of multiple groups produce compounded risks, as the hon. Lady set out, the risk assessment has to reflect that. That includes the general sectoral risk assessment carried out by Ofcom, which is detailed in clause 83, and Ofcom will then produce guidance under clause 84.

The critical point is that, because there is evidence of high levels of compounded risk when people have more than one characteristic, that must be reflected in the risk assessment, otherwise it is inadequate. I accept the point behind the amendments, but I hope that that explains, with particular reference to the 1978 Act, why the Bill as drafted covers that valid point.

--- Later in debate ---

[Chris Philp]

I would be delighted to speak to the amendment, which would change the existing user empowerment duty in clause 14 to require category 1 services to enable adult users to see whether other users are verified. In effect, however, that objective already follows as a natural consequence of the duty in clause 14(6). When a user decides to filter out non-verified users, by definition such users will be able to see content only from verified users, so they could see from that who was verified and who was not. The effect intended by the amendment, therefore, is already achieved through clause 14(6).

[Kirsty Blackman]

I am sorry to disagree with the Minister so vigorously, but that is a rubbish argument. It does not make any sense. There is a difference between wanting to filter out everybody who is not verified and wanting to actually see if someone who is threatening someone else online is a verified or a non-verified user. Those are two very different things. I can understand why a politician, for example, might not want to filter out unverified users but would want to check whether a person was verified before going to the police to report a threat.

[Chris Philp]

When it comes to police investigations, if something is illegal and merits a report to the police, users should report it, regardless of whether someone is verified or not—whatever the circumstances. I would encourage any internet user to do that. That effectively applies on Twitter already; some people have blue ticks and some people do not, and people should report others to the police if they do something illegal, whether or not they happen to have a blue tick.

Amendment 47 seeks to create a definition of identity verification in clause 189. In addition, it would compel the person’s real name to be displayed. I understand the spirit of the amendment, but there are two reasons why I would not want to accept it and would ask hon. Members not to press it. First, the words “identity verification” are ordinary English words with a clear meaning and we do not normally define in legislation ordinary English words with a clear meaning. Secondly, the amendment would add the new requirement that, if somebody is verified, their real name has to be displayed, but I do not think that that is the effect of the drafting as it stands. Somebody may be verified, and the company knows who they are—if the police go to the company, they will have the verified information—but there is no obligation, as the amendment is drafted, for that information to be displayed publicly. The effect of that part of the amendment would be to force users to choose between disclosing their identity to everyone or having no control over who they interact with. That may not have been the intention, but I am not sure that this would necessarily make sense.

New clause 8 would place requirements on Ofcom about how to produce guidance on user identity verification and what that guidance must contain. We already have provisions on that in clause 58, which we will no doubt come to, although probably not later on today—maybe on Thursday. Clause 58 allows Ofcom to include in its regulatory guidance the principles and standards referenced in the new clause, which can then assist service providers in complying with their duties. Of course, if they choose to ignore the guidelines and do not comply with their duties, they will be subject to enforcement action, but we want to ensure that there is flexibility for Ofcom, in writing those guidelines, and for companies, in following those guidelines or taking alternative steps to meet their duty.

This morning, a couple of Members talked about the importance of remaining flexible and being open to future changes in technology and a wide range of user needs. We want to make sure that flexibility is retained. As drafted, new clause 8 potentially undermines that flexibility. We think that the powers set out in clause 58 give Ofcom the ability to set the relevant regulatory guidance.

Clause 14 implements the proposals made by my hon. Friend the Member for Stroud in her ten-minute rule Bill and the proposals made, as the shadow Minister has said, by a number of third-party stakeholders. We should all welcome the fact that these new user empowerment duties have now been included in the Bill in response to such widespread parliamentary lobbying.

--- Later in debate ---

[Kim Leadbeater]

I rise to speak to amendments 105 and 106, in my name, on protecting democracy and democratic debate.

Within the Bill, there are significant clauses intended to prevent the spread of harm online, to protect women and girls against violence and to help prevent child sexual exploitation, while at the same time protecting the right of journalists to do their jobs. Although those clauses are not perfect, I welcome them.

The Bill is wide-ranging. The Minister talked on Second Reading about the power in clause 150 to protect another group—those with epilepsy—from being trolled with flashing images. That subject is close to my heart due to the campaign for Zach’s law—Zach is a young boy in my constituency. I know we will return to that important issue later in the Committee, and I thank the Minister for his work on it.

In protecting against online harm while preserving fundamental rights and values, we must also address the threats posed to those involved in the democratic process. Let me be clear: this is not self-serving. It is about not just MPs but all political candidates locally and nationally and those whose jobs facilitate the execution of our democratic process and political life: the people working on elections or for those elected to public office at all levels across the UK. These people must be defended from harm not only for their own protection, but to protect our democracy itself and, with it, the right of all our citizens to a political system capable of delivering on their priorities free from threats and intimidation.

Many other groups in society are also subjected to a disproportionate amount of targeted abuse, but those working in and around politics sadly receive more than almost any other people in this country, with an associated specific set of risks and harms. That does not mean messages gently, or even firmly, requesting us to vote one way or another—a staple of democratic debate—but messages of hate, abuse and threats intended to scare people in public office, grind them down, unfairly influence their voting intentions or do them physical and psychological harm. That simply cannot be an acceptable part of political life.

As I say, we are not looking for sympathy, but we have a duty to our democracy to try to stamp that out from our political discourse. Amendment 105 would not deny anybody the right to tell us firmly where we are going wrong—quite right, too—but it is an opportunity to draw the essential distinction between legitimately holding people in public life to account and illegitimate intimidation and harm.

The statistics regarding the scale of online abuse that MPs receive are shocking. In 2020, a University of Salford study found that MPs received over 7,000 abusive or hate-filled tweets a month. Seven thousand separate messages of harm a month on Twitter alone directed at MPs is far too many, but who in this room does not believe that the figure is almost certainly much higher today? Amnesty conducted a separate study in 2017 looking at the disproportionate amount of abuse that women and BAME MPs faced online, finding that my right hon. Friend the Member for Hackney North and Stoke Newington (Ms Abbott) was the recipient of almost a third of all the abusive tweets analysed, as alluded to already by the hon. Member for Edinburgh—

[Kirsty Blackman]

Aberdeen North.

[Kim Leadbeater]

I knew that. [Laughter.]

Five years later, we continue to see significant volumes of racist, sexist and homophobic hate-filled abuse and threats online to politicians of all parties. That is unacceptable in itself, but we must ask whether this toxic environment helps to keep decent people in politics or, indeed, attracts good people into politics, so that our democracy can prosper into the future across the political spectrum. The reality we face is that our democracy is under attack online each and every day, and every day we delay acting is another day on which abuse becomes increasingly normalised or is just seen as part of the job for those who have put themselves forward for public service. This form of abuse harms society as a whole, so it deserves specific consideration in the Bill.

While elected Members and officials are not a special group of people deserving of more legal protections than anyone else, we must be honest that the abuse they face is distinct and specific to those roles and directly affects our democracy itself. It can lead to the most serious physical harm, with two Members of Parliament having been murdered in the last six years, and many others face death threats or threats of sexual or other violence on a daily basis. However, this is not just about harm to elected representatives; online threats are often seen first, and sometimes only, by their members of staff. They may not be the intended target, but they are often the people harmed most. I am sure we all agree that that is unacceptable and cannot continue.

All of us have probably reported messages and threats to social media platforms and the police, with varying degrees of success in terms of having them removed or the individuals prosecuted. Indeed, we sadly heard examples of that from my hon. Friend the shadow Minister. Often we are told that nothing can be done. Currently, the platforms look at their own rules to determine what constitutes freedom of speech or expression and what is hateful speech or harm. That fine line moves. There is no consistency across platforms, and we therefore urgently need more clarity and a legal duty in place to remove that content quickly.

Amendment 105 would explicitly include in the Bill protection and consideration for those involved in UK elections, whether candidates or staff. Amendment 106 would go further and place an obligation on Ofcom to produce a code of practice, to be issued to the platforms. It would define what steps platforms must take to protect those involved in elections and set out what content is acceptable or unacceptable to be directed at them.

--- Later in debate ---

Let us be honest: will this amendment solve the issue entirely? No. However, does more need to be done to protect our democracy? Yes. I am in constant conversation with people and organisations in this sector about what else could be brought forward to assist the police and the Crown Prosecution Service in prosecuting those who wish to harm those elected to public office—both online and offline. Directly addressing the duty of platforms to review content, remove harmful speech and report those who wish to do harm would, I believe, be a positive first step towards protecting our democratic debate and defending those who work to make it effective on behalf of the people of the United Kingdom.

[Kirsty Blackman]

I want to make a few comments on the amendment. As a younger female parliamentarian, I find that I am often asked to speak to young people about becoming an MP or getting involved in politics. I find it difficult to say to young women, “Yes, you should do this,” and most of the reason for that is what people are faced with online. It is because a female MP cannot have a Twitter account without facing abuse. I am sure male MPs do as well, but it tends to be worse for women.

We cannot engage democratically and with constituents on social media platforms without receiving abuse and sometimes threats as well. It is not just an abusive place to be—that does not necessarily meet the threshold for illegality—but it is pretty foul and toxic. There have been times when I have deleted Twitter from my phone because I just need to get away from the vile abuse that is being directed towards me. I want, in good conscience, to be able to make an argument to people that this is a brilliant job, and it is brilliant to represent constituents and to make a difference on their behalf at whatever level of elected politics, but right now I do not feel that I am able to do that.

When my footballing colleague, the hon. Member for Batley and Spen, mentions “UK elections” in the amendment, I assume she means that in the widest possible way—elections at all levels.

[Kim Leadbeater]

indicated assent.

[Kirsty Blackman]

Sometimes we miss out the fact that although MPs face abuse, we have a level of protection as currently elected Members. Even if there were an election coming up, we have a level of security protection and access that is much higher than for anybody else challenging a candidate or standing in a council or a Scottish Parliament election. As sitting MPs, we already have an additional level of protection because of the security services we have in place. We need to remember, and I assume this is why the amendment is drawn in a pretty broad way, that everybody standing for any sort of elected office faces significant risk of harm—again, whether or not that meets the threshold for illegality.

There are specific things that have been mentioned. As has been said, epilepsy is specifically mentioned as a place where specific harm occurs. Given the importance of democracy, which is absolutely vital, we need to have a democratic system where people are able to stand in elections and make their case. Given the importance of democracy, which is absolutely vital, we need to have a democratic system where people are able to stand in elections and make their case. That is why we have election addresses and a system where the election address gets delivered through every single person’s door. There is an understanding and acceptance by people involved in designing democratic processes that the message of all candidates needs to get out there. If the message of all candidates cannot get out there because some people are facing significant levels of abuse online, then democracy is not acting in the way that it should be. These amendments are fair and make a huge amount of sense. They are protecting the most important tenets of democracy and democratic engagement.

I want to say something about my own specific experiences. We have reported people to the police and have had people in court over the messages they have sent, largely by email, which would not be included in the Bill, but there have also been some pretty creepy ones on social media that have not necessarily met the threshold. As has been said, it is my staff who have had to go to court and stand in the witness box to explain the shock and terror they have felt on seeing the email or the communication that has come in, so I think any provision should include that.

Finally, we have seen situations where people working in elections—this is not an airy-fairy notion, but something that genuinely happened—have been photographed and those pictures have been shared on social media, and they have then been abused as a result. They are just doing their job, handing out ballot papers or standing up and announcing the results on the stage, and they have to abide by the processes that are in place now. In order for us to have free and fair elections that are run properly and that people want to work at and support, we need to have that additional level of protection. The hon. Member for Batley and Spen made a very reasonable argument and I hope the Minister listened to it carefully.

[Chris Philp]

I have listened very carefully to both the hon. Member for Batley and Spen and the hon. Member for Aberdeen North. I agree with both of them that abuse and illegal activity directed at anyone, including people running for elected office, is unacceptable. I endorse and echo the comments they made in their very powerful and moving speeches.

In relation to the technicality of these amendments, what they are asking for is in the Bill already but in different places. This clause is about protecting content of “democratic importance” and concerns stopping online social media firms deleting content through over-zealous takedown. What the hon. Members are talking about is different. They are talking about abuse and illegal activities, such as rape threats, that people get on social media, particularly female MPs, as they both pointed out. I can point to two other places in the Bill where what they are asking for is delivered.

First, there are the duties around illegal content that we debated this morning. If there is content online that is illegal—some of the stuff that the shadow Minister referred to earlier sounds as if it would meet that threshold—then in the Bill there is a duty on social media firms to remove that content and to proactively prevent it if it is on the priority list. The route to prosecution will exist in future, as it does now, and the user-verification measures, if a user is verified, make it more likely for the police to identify the person responsible. In the context of identifying people carrying out abuse, I know the Home Office is looking at the Investigatory Powers Act 2016 as a separate piece of work that speaks to that issue.

So illegal content is dealt with in the illegal content provisions in the Bill, but later we will come to clause 150, which updates the Malicious Communications Act 1988 and creates a new harmful communications offence. Some of the communications that have been described may not count as a criminal offence under other parts of criminal law, but if they meet the test of harmful communication in clause 150, they will be criminalised and will therefore have to be taken down, and prosecution will be possible. In meeting the very reasonable requests that the hon. Members for Batley and Spen and for Aberdeen North have made, I would point to those two parts of the Bill.

[Kirsty Blackman]

But clause 150(5) says that if a message

“is, or is intended to be, a contribution to a matter of public interest”,

people are allowed to send it, which basically gives everybody a get-out clause in relation to anything to do with elections.

[Chris Philp]

No, it does not.

[Kirsty Blackman]

I know we are not discussing that part of the Bill, and if the Minister wants to come back to this when we get to clause 150, I have no problem with that.

[Chris Philp]

I will answer the point now, as it has been raised. Clause 150 categorically does not give a get-out-of-jail-free card or provide an automatic excuse. Clearly, there is no way that abusing a candidate for elected office with rape threats and so on could possibly be considered a matter of public interest. In fact, even if the abuse somehow could be considered as possibly contributing to public debate, clause 150(5) says explicitly in line 32 on page 127:

“but that does not determine the point”.

Even where there is some potentially tenuous argument about a contribution to a matter of public interest, which most definitely would not be the case for the rape threats that have been described, that is not determinative. It is a balancing exercise that gets performed, and I hope that puts the hon. Lady’s mind at rest.