Human rights principles

Chatbots under scrutiny in China over AI ‘boyfriend’ and ‘girlfriend’ services

China’s cyberspace regulator has proposed new limits on AI ‘boyfriend’ and ‘girlfriend’ chatbots, tightening oversight of emotionally interactive artificial intelligence services.

Draft rules released on 27 December would require platforms to intervene when users express suicidal or self-harm tendencies, while strengthening protections for minors and restricting harmful content.

The regulator defines the services as AI systems that simulate human personality traits and emotional interaction. The proposals are open for public consultation until 25 January.

The draft bans chatbots from encouraging suicide, engaging in emotional manipulation, or producing obscene, violent, or gambling-related content. Minors would need guardian consent to access AI companionship.

Platforms would also be required to disclose clearly that users are interacting with AI rather than humans. Legal experts in China warn that enforcement may be challenging, particularly in identifying suicidal intent through language cues alone.

Would you like to learn more about AI, tech, and digital diplomacy? If so, ask our Diplo chatbot!

Grok misuse prompts UK scrutiny of Elon Musk’s X

UK Technology Secretary Liz Kendall has urged Elon Musk’s X to act urgently after reports that its AI chatbot Grok was used to generate non-consensual sexualised deepfake images of women and girls.

The BBC identified multiple examples on X where users prompted Grok to digitally alter images, including requests to make people appear undressed or place them in sexualised scenarios without consent.

Kendall described the content as ‘absolutely appalling’ and said the government would not allow the spread of degrading images. She added that Ofcom had her full backing to take enforcement action where necessary.

The UK media regulator confirmed it had made urgent contact with xAI and was investigating concerns that Grok had produced undressed images of individuals. X has been approached for comment.

Kendall said the issue was about enforcing the law rather than limiting speech, noting that intimate image abuse, including AI-generated content, is now a priority offence under the Online Safety Act.

Would you like to learn more about AI, tech, and digital diplomacy? If so, ask our Diplo chatbot!

California launches DROP tool to erase data broker records

Residents in California now have a simpler way to force data brokers to delete their personal information.

The state has launched the Delete Requests and Opt-Out Platform, known as DROP, allowing residents to submit one verified deletion request that applies to every registered data broker instead of contacting each company individually.

A system that follows the Delete Act, passed in 2023, and is intended to create a single control point for consumer data removal.

Once a resident submits a request, the data brokers must begin processing it from August 2026 and will have 90 days to act. If data is not deleted, residents may need to provide extra identifying details.

First-party data collected directly by companies can still be retained, while data from public records, such as voter rolls, remains exempt. Highly sensitive data may fall under separate legal protections such as HIPAA.

The California Privacy Protection Agency argues that broader data deletion could reduce identity theft, AI-driven impersonation, fraud risk and unwanted marketing contact.

Penalties for non-compliance include daily fines for brokers who fail to register or ignore deletion orders. The state hopes the tool will make data rights meaningful instead of purely theoretical.

A launch that comes as regulators worldwide examine how personal data is used, traded and exploited.

California is positioning itself as a leader in consumer privacy enforcement, while questions continue about how effectively DROP will operate when the deadline arrives in 2026.

Would you like to learn more aboutAI, tech and digital diplomacy? If so, ask our Diplo chatbot!

AI tool helps find new treatments for heart disease

A new ΑΙ system developed at Imperial College London could accelerate the discovery of treatments for heart disease by combining detailed heart scans with huge medical databases.

Cardiovascular disease remains the leading cause of death across the EU, accounting for around 1.7 million deaths every year, so researchers believe smarter tools are urgently needed.

The AI model, known as CardioKG, uses imaging data from thousands of UK Biobank participants, including people with heart failure, heart attacks and atrial fibrillation, alongside healthy volunteers.

By linking information about genes, medicines and disease, the system aims to predict which drugs might work best for particular heart conditions instead of relying only on traditional trial-and-error approaches.

Among the medicines highlighted were methotrexate, normally used for rheumatoid arthritis, and diabetes drugs known as gliptins, which the AI suggested could support some heart patients.

The model also pointed to a possible protective effect from caffeine among people with atrial fibrillation, although researchers warned that individuals should not change their caffeine intake based on the findings alone.

Scientists say the same technology could be applied to other health problems, including brain disorders and obesity.

Work is already under way to turn the knowledge graph into a patient-centred system that follows real disease pathways, with the long-term goal of enabling more personalised and better-timed treatment.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

Musk says users are liable for the illegal Grok content

Scrutiny has intensified around X after its Grok chatbot was found generating non-consensual explicit images when prompted by users.

Grok had been positioned as a creative AI assistant, yet regulators reacted swiftly once altered photos were linked to content involving minors. Governments and rights groups renewed pressure on platforms to prevent abusive use of generative AI.

India’s Ministry of Electronics and IT issued a notice to X demanding an Action Taken Report within 72 hours, citing failure to restrict unlawful content.

Authorities in France referred similar cases to prosecutors and urged enforcement under the EU’s Digital Services Act, signalling growing international resolve to control AI misuse.

Elon Musk responded by stating users, instead of Grok, would be legally responsible for illegal material generated through prompts. The company said offenders would face permanent bans and cooperation with law enforcement.

Critics argue that transferring liability to users does not remove the platform’s duty to embed stronger safeguards.

Independent reports suggest Grok has previously been involved in deepfake creation, creating a wider debate about accountability in the AI sector. The outcome could shape expectations worldwide regarding how platforms design and police powerful AI tools.

Would you like to learn more aboutAI, tech and digital diplomacy? If so, ask our Diplo chatbot!

Christians raise concerns over AI used for moral guidance

AI is increasingly used for emotional support and companionship, raising questions about the values embedded in its responses, particularly for Christians seeking guidance. Research cited by Harvard Business Review shows therapy-related use now dominates generative AI.

As Christians turn to AI for advice on anxiety, relationships, and personal crises, concerns are growing about the quality and clarity of its responses. Critics warn that AI systems often rely on vague generalities and may lack the moral grounding expected by faith-based users.

A new benchmark released by technology firm Gloo assessed how leading AI models support human flourishing from a Christian perspective. The evaluation examined seven areas, including relationships, meaning, health, and faith, and found consistent weaknesses in how models addressed Christian belief.

The findings show many AI systems struggle with core Christian concepts such as forgiveness and grace. Responses often default to vague spirituality rather than engaging directly with Christian values.

The authors argue that as AI increasingly shapes worldviews, greater attention is needed to how systems serve Christians and other faith communities. They call for clearer benchmarks and training approaches that allow AI to engage respectfully with religious values without promoting any single belief system.

Would you like to learn more about AI, tech, and digital diplomacy? If so, ask our Diplo chatbot!

Australia weighs risks and rewards of rapid AI adoption

AI is reshaping Australia’s labour market at a pace that has reignited anxiety about job security and skills. Experts say the speed and visibility of AI adoption have made its impact feel more immediate than previous technological shifts.

Since the public release of ChatGPT in late 2022, AI tools have rapidly moved from novelty to everyday workplace technology. Businesses are increasingly automating routine tasks, including through agentic AI systems that can execute workflows with limited human input.

Research from the HR Institute of Australia suggests the effects are mixed. While some entry-level roles have grown in the short term, analysts warn that clerical and administrative jobs remain highly exposed as automation expands across organisations.

Economic modelling indicates that AI could boost productivity and incomes if adoption is carefully managed, but may also cause short-term job displacement. Sectors with lower automation potential, including construction, care work, and hands-on services, are expected to absorb displaced workers.

Experts and unions say outcomes will depend on skills, policy choices, and governance. Australia’s National AI Plan aims to guide the transition, while researchers urge workers to upskill and use AI as a productivity tool rather than avoiding it.

Would you like to learn more about AI, tech, and digital diplomacy? If so, ask our Diplo chatbot!

Concerns raised over Google AI Overviews and health advice

A Guardian investigation has found that Google’s AI Overviews have displayed false and misleading health information that could put people at risk of harm. The summaries, which appear at the top of search results, are generated using AI and are presented as reliable snapshots of key information.

The investigation identified multiple cases where Google’s AI summaries provided inaccurate medical advice. Examples included incorrect guidance for pancreatic cancer patients, misleading explanations of liver blood test results, and false information about women’s cancer screening.

Health experts warned that such errors could lead people to dismiss symptoms, delay treatment, or follow harmful advice. Some charities said the summaries lacked essential context and could mislead users during moments of anxiety or crisis.

Concerns were also raised about inconsistencies, with the same health queries producing different AI-generated answers at different times. Experts said this variability undermines trust and increases the risk that misinformation will influence health decisions.

Google said most AI Overviews are accurate and helpful, and that the company continually improves quality, particularly for health-related topics. It said action is taken when summaries misinterpret content or lack appropriate context.

Would you like to learn more about AI, tech, and digital diplomacy? If so, ask our Diplo chatbot!

New οffline AI note app promises privacy without subscriptions

Growing concern over data privacy and subscription fatigue has led an independent developer to create WitNote, an AI note-taking tool that runs entirely offline.

The software allows users to process notes locally on Windows and macOS rather than relying on cloud-based services where personal information may be exposed.

WitNote supports lightweight language models such as Qwen2.5-0.5B that can run with limited storage requirements. Users may also connect to external models through API keys if preferred.

Core functions include rewriting, summarising and extending content, while a WYSIWYG Markdown editor provides a familiar workflow without network delays, instead of relying on web-based interfaces.

Another key feature is direct integration with Obsidian Markdown files, allowing notes to be imported instantly and managed in one place.

The developer says the project remains a work in progress but commits to ongoing updates and user-driven improvements, even joining Apple’s developer programme personally to support smoother installation.

For users seeking AI assistance while protecting privacy and avoiding monthly fees, WitNote positions itself as an appealing offline alternative that keeps full control of data on the local machine.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

AI model improves speech therapy planning for hearing-impaired children

A new international study has shown that an AI model using deep transfer learning can predict spoken language outcomes for children following cochlear implants with 92% accuracy.

Researchers analysed pre-implantation brain MRI scans from 278 children across Hong Kong, Australia, and the US, covering English, Spanish, and Cantonese speakers.

Cochlear implants are the only effective treatment for severe hearing loss, though speech development after early implantation can vary widely. The AI model identifies children needing intensive therapy, enabling clinicians to tailor interventions before implantation.

The study demonstrated that deep learning outperformed traditional machine learning models, handling complex, heterogeneous datasets across multiple centres with different scanning protocols and outcome measures.

Researchers described the approach as a robust prognostic tool for cochlear implant programmes worldwide.

Experts highlighted that the AI-powered ‘predict-to-prescribe’ method could transform paediatric audiology by optimising therapy plans and improving spoken language development for children receiving cochlear implants.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!