UK targets risky AI chatbots with fines, bans

The UK plans to introduce new regulations targeting AI chatbots that may pose risks to children, with potential fines or service bans.

Keir Starmer is set to announce measures to address illegal content created by AI, following public concerns over platforms like Elon Musk’s X halting its Grok AI tool from generating sexualized images of real people in the UK.

The government aims to close legal loopholes under the Online Safety Act, requiring AI chatbot providers to adhere to content restrictions or face penalties, including fines up to 10% of global revenue or court-ordered bans.

The move follows the UK regulator Ofcom’s admission that current laws do not sufficiently cover AI-generated images and videos, unless classified as pornography.

Critics, including the NSPCC, warn that AI chatbots have been linked to harmful content for minors, such as misinformation about body image and self-harm.

The government plans to consult on further restrictions for social media use by under-16s, potentially as early as this summer.

🔗 Source: The Guardian

🧠 Food for thought

Implications, context, and why it matters.

The UK’s crackdown follows evidence of risks across the AI industry

Action against Grok follows a joint investigation that estimated the tool produced 3 million sexualised images in 11 days, including 23,000 that appear to depict children 1.
A study of consumer chatbots, including OpenAI’s ChatGPT and Google’s Gemini, found the systems could estimate users’ ages. Even so, none blocked identified child users in 1,050 tests, including 720 child-age-group tests 2.
UK ministers and regulator Ofcom said some AI chatbots may sit outside parts of the Online Safety Act. Coverage can depend on use such as search services, user-to-user features, or generating pornography 3.

A push for age assurance could change AI business models

UK rules are speeding up a shift toward stronger age assurance (ways to check a user’s age) for AI services.
OpenAI has signalled more age checks in ChatGPT. Some users may face prompts to complete verification steps through third-party vendors, tied to a wider effort to separate minors from adults 4.
Verification can slow sign-ups and raise privacy concerns when accounts link to real-world identities 4.
The Online Safety Act can cover services with UK links even when providers operate abroad, which may lead companies to add compliance features that reach non-UK users 5.