Online Safety Act Network

ChatGPT found to be giving teens 'dangerous' advice on mental health More than half of ChatGPT's 1,200 responses were found to be "dangerous", according to a study by The Center for Countering Digital Hate

ChatGPT is betraying vulnerable teenagers.

We tested OpenAI’s platform & found it generated harmful answers to over HALF of 1.2K prompts.

AI companies must take safety seriously. CCDH's new research in @themirror.com 👇

Explainer here on some of the issues that have dominated the news in recent weeks. ⬇️
www.onlinesafetyact.net/analysis/the...

Explainer here on some of the issues that have dominated the news in recent weeks. ⬇️
www.onlinesafetyact.net/analysis/the...

📣 Online safety campaigners, free speech advocates, and tech experts

If you have thoughts and insights on the Online Safety Act, I’d love to hear them

💬 Drop me a DM/comment below

Ruling here: www.judiciary.uk/judgments/wi...

Ruling here: www.judiciary.uk/judgments/wi...

#onlinesafetyact JR ruling due today | via @politico.com

Blame Pornhub for Britain’s Age-Check Internet Mess The backlash against the UK’s new child safety law should also go to companies who built glitchy age-checking systems.

"Far from being a policy miscalculation, [the OSA] is doing exactly what it's supposed to, and part of the reason for its painful transition is that platforms have dragged their feet on building technically seamless age-checking systems." | @parmy.bsky.social www.bloomberg.com/opinion/arti...

Keeping children safe online: changes to the Online Safety Act explained How new laws that keep children safe on the internet work.

Useful explainer from #DSIT ⬇️ #onlinesafetyact

www.gov.uk/government/n...

This. ⬇️ #onlinesafetyact

Something for the weekend. 😉 #onlinesafetyact www.onlinesafetyact.net/analysis/the...

If anything, the Online Safety Act doesn't go far enough It’s deeply concerning to see the rhetoric around the Online Safety Act shift toward loss of free expression, writes the NSPCC chief executive

Amid all the noise of the last week, this is a vital reminder from Chris Sherwood @nspcc.bsky.social as to why we have the #onlinesafetyact, why parents and campaigners backed it and why it had broad cross-party support in Parliament when it passed. ⬇️ www.politicshome.com/opinion/arti...

💬 “The law is a good start, but we must keep speaking up for a digital world made for us.”

🇬🇧 Today, the Children’s Code under the Online Safety Act comes into effect, holding tech companies accountable.

❓ What do young people think?

Read more: https://loom.ly/kzxwdCs

#OnlineSafetyAct

Study: Social media boosts climate conspiracies over emergency alerts The study says the influence of high-profile conspiracy theorists during climate disasters is drowning out emergency response efforts.

🚨 CCDH's new research shows that conspiracies get more views than official information during extreme weather events, like the LA wildfires.

X, Meta & YouTube must stop enabling and profiting from extreme weather false claims.

Read more in @euronews.com ⤵️

Media regulator contacts X over age verification concerns The media regulator Coimisiún na Meán has contacted X, formerly Twitter, amid concerns over its age verification measures.

Age-verification requirements went into effect Monday in Ireland. X appears to be in violation, says the media regulator, and could be on the hook for 10% of overall revenue if it fails to block access to pornography and Tide Pod challenges.

www.rte.ie/news/2025/07...

The #onlinesafetyact children's safety duties come into force on Friday. Here's what to expect. ⬇️

www.onlinesafetyact.net/analysis/the...

CEO of Elon Musk’s X, Linda Yaccarino, resigns The announcement comes a day after the social media site’s AI chatbot made antisemitic statements. Yaccarino did not offer a reason for her resignation.

Linda Yaccarino, CEO of Elon Musk’s X, announced Wednesday that she is stepping down from the social media platform after two years in the position. www.washingtonpost.com/technology/2...

Parliamentlive.tv Speaker's Conference (2024)

Hearing due to start imminently: parliamentlive.tv/Event/Index/...

I cannot think of a reason why the regulators shouldn't do this. There's a clear pattern of negligent behaviour in the long lead up, followed by what happened yesterday. They are unable to operate a safe platform.

Speaker’s Conference to hear evidence from X and Meta - Committees - UK Parliament The Speaker’s Conference (2024) will on Wednesday, 9th July 2025 hear from X (formerly Twitter), Meta and a senior lecturer from the University of Glasgow.

Hauling in X, you say? committees.parliament.uk/committee/74...

Researchers’ access to information from regulated online services Ofcom has today published a report examining how researchers could gain better access to information from regulated online services to support work on online safety matters. The report outlines three ...

@ofcom.bsky.social has published its report on researcher access to information - a critical part of the #onlinesafety regime. (Provisions for the researcher access framework were included in the recently-passed Data (Use and Access) Act.) #onlinesafetyact www.ofcom.org.uk/online-safet...

NEW: check out this excellent explainer on #chatbots and the #onlinesafetyact from our expert legal adviser, Prof Lorna Woods OBE ⬇️ #AI #artificialintelligence #regulation #onlineharms #onlinesafety #socialmedia #internetregulation www.onlinesafetyact.net/analysis/cha...

Important update from @ofcom.bsky.social , with very welcome signs of compliance from the major porn providers in the UK with #onlinesafetyact duties to prevent under-18s accessing porn ahead of the enforcement deadline next month. ⬇️ www.ofcom.org.uk/online-safet...

AI Chatbots Are Making LA Protest Disinformation Worse Amid fast-moving events in Los Angeles, users are turning to chatbots like Grok and ChatGPT to find out what’s real and what’s not—and getting inaccurate information.

Tech companies' abandoning fact-checking on their platforms has led to disinformation about the protests against the #ICERaids in LA to go viral.

And AI-powered chatbots like Grok and ChatGPT are making the situation considerably worse

www.wired.com/story/grok-c...

'Nine new online safety investigations into 4chan, file-sharing services, and a porn provider'

🚨 9 new online safety investigations.

We're investigating whether 4chan, a porn provider and seven file-sharing services have failed to comply with their duties under the Online Safety Act.

🔗 Find out more: www.ofcom.org.uk/online-safet...

UK campaigners raise alarm over report of Meta plan to use automation for risk checks Ofcom ‘considering the concerns’ raised after claim that up to 90% of risk assessments will be carried out by AI

Meta wants 90% of its risk assessments done by AI, including on child safety & misinformation.

CCDH joined leading organisations urging Ofcom to limit AI use in risk assessments, as it undermines the spirit of the Online Safety Act. More in @theguardian.com: www.theguardian.com/technology/2...

YouTube Loosens Video Content Moderation Rules The world’s largest video platform has told content moderators to favor “freedom of expression” over the risk of harm in deciding what to take down.

“What we’re seeing is a rapid race to the bottom. The changes benefit the companies by reducing the costs of content moderation, while keeping more content online..."

Our CEO, @imranahmed.bsky.social in @nytimes.com on YouTube's choice to loosen content moderation to allow more harmful content 👇

Here's the Government response to the responses received during the consultation. 3/ assets.publishing.service.gov.uk/media/684198...

Here's a link to the draft regulations themselves. These now require Parliamentary approval. They also don't come into force until 31st December 2025 as @ofcom.bsky.social is required to consult on guidance to support orgs who might submit super-complaints. 2/ www.legislation.gov.uk/ukdsi/2025/9...

Written statements - Written questions, answers and statements - UK Parliament Information from UK Parliament on written questions & answers, written statements and daily reports.

🚨The wait is over. #DSIT has *finally* got the draft regulations for the #onlinesafetyact #supercomplaints mechanism out into the world (the consultation on the regime closed in January 2024). Here's the WMS announcing their laying in Parliament questions-statements.parliament.uk/written-stat... 1/