Online Safety Act 2023: Repeal

Manuela Perteghella Excerpts

Monday 15th December 2025 Westminster Hall

[Lewis Atkinson]

My hon. Friend is right. The proactive duty that the Act places on providers in relation to the nature of their algorithms and their content is crucial because of the type of content to which she refers. It is right that the largest providers, and those most frequently used by kids, have to take active responsibility for keeping children safe. The implementation of the OSA means that algorithms serving harmful content to kids are now being regulated for the first time. There is a long way to go, and I am sure that other Members will say more than I can in this introduction, but I want to be clear to my constituents that I support the action that the OSA is prompting to improve children’s safety and welfare online.

Various surveys set out the impact of the Online Safety Act; Ofcom is publishing its research and a formal Government review will follow in due course. However, most impactful for me was seeing a teenage boy say on a news piece recently that, now,

“when I’m scrolling TikTok, I’m free from violence.”

That changed for him in the months following the implementation of the Online Safety Act, so it is no wonder that organisations such as the Online Safety Act Network, which I spoke to in preparation for this debate, fully support the Act’s principles. The network points to early evidence that the Act is actively reducing harm to children and emphasised that Ofcom must move beyond content filters to ensure safety by design, which would, for example, include addressing features that incentivise pile-ons, targeting an individual with abuse and harassment.

New Ofcom research shows that 58% of parents now believe that measures in the code of practice are beginning to improve the safety of children online. My belief is that we should be considering not whether to repeal the Act, but how we can continue to enforce it in a robust, effective and proportionate manner.

The way in which the Online Safety Act addresses online hate has perhaps not had as much focus as it might have. As well as being a member of the Petitions Committee, I am privileged to be a member of the Home Affairs Committee, which is conducting an inquiry into combating new forms of extremism. It is very clear from the public evidence that we have received so far that, left unregulated and unchallenged, online spaces and services can be used to amplify hate, thus risking a rise in extremist action, including violence.

Analysis by the Antisemitism Policy Trust highlights that there are patterns of co-ordinated and persistent misogynistic, anti-immigrant, anti-Government and antisemitic discourse on social media, with bot accounts being repeatedly used to amplify misleading or harmful narratives that fuel hate and may increase the risk of violence. Such content often breaches platforms’ own terms of service, but under the Online Safety Act, I understand that Ofcom category 1 services will now be mandated to proactively offer users optional tools to help them to reduce the likelihood that they will encounter legal but harmful content such as that.

There is much to be done to implement those provisions in an appropriate manner. However, I invite anyone calling for full repeal of the Act to consider how we as a society deal with the rise of extremism, and a context where the internet can be used as a sort of free-for-all fuelled by hate-filled algorithms that thrive on and incentivise division and hatred, rather than consensus and civic peace.

I am aware that there are large parts of the Online Safety Act that I have not been able to touch on today; I hope that others will do so during the debate. There are questions about end-to-end encryption, cyber-flashing, the creation of abusive deepfakes, AI moderation and chatbots.

[Manuela Perteghella (Stratford-on-Avon) (LD)]

The hon. Member is making a strong and thoughtful case. Does he agree that although the Act regulates user-to-user services, it leaves a significant gap around generative AI chatbots, despite the growing evidence of harm caused to children from private interaction with them? And does he share my concern that the speed at which this technology is developing risks outpacing the legislative framework that we have in place?

[Lewis Atkinson]

I agree with the hon. Lady. In my understanding, when the legislation was drafted, it was not initially clear to those who drafted it that AI would develop at the astonishing pace that it has in recent years. I ask the Minister to reflect on that point in addressing the implementation of the Act and its potential future evolution through primary legislation.