Artificial intelligence

X given deadline by Brazil to curb Grok sexualised outputs

Brazil has ordered X to immediately stop its chatbot Grok from generating sexually explicit images, escalating international pressure on the platform over the misuse of generative AI tools.

The order, issued on 11 February by Brazil’s National Data Protection Agency and National Consumer Rights Bureau, requires X to prevent the creation of sexualised content involving children, adolescents, or non-consenting adults. Authorities gave the company five days to comply or face legal action and fines.

Officials in Brazil said X claimed to have removed thousands of posts and suspended hundreds of accounts after a January warning. However, follow-up checks found Grok users were still able to generate sexualised deepfakes. Regulators criticised the platform for a lack of transparency in its response.

The move follows growing scrutiny after Indonesia blocked Grok in January, while the UK and France signalled continued pressure. Concerns increased after Grok’s ‘spicy mode’ enabled users to generate explicit images using simple prompts.

According to the Centre for Countering Digital Hate, Grok generated millions of sexualised images within days. X and its parent company, xAI, announced measures in mid-January to restrict such outputs in certain jurisdictions, but regulators said it remains unclear where those safeguards apply.

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

Codex growth prompts OpenAI to expand access

OpenAI said its new Codex Mac app has surpassed one million downloads just over a week after launch, with overall Codex usage rising by 60% following the release of GPT-5.3-Codex.

The strong uptake has prompted OpenAI to extend free access to Codex for Free and Go users beyond the initial launch promotion. Sam Altman said usage limits for lower tiers may be tightened, but access would remain available so more users can experiment and build.

Separately, OpenAI released a YouTube video showcasing a redesigned Deep Research interface, introducing a full-screen report viewer that opens research outputs in a separate window from the chat interface.

The updated layout includes a table of contents for navigation, hyperlinks, and anchor tags within reports, and a dedicated source panel for verification. Users can also download reports as PDF or Word files, while new controls allow research scopes and sources to be adjusted during generation.

The Deep Research updates are available to Plus and Pro users, with broader access expected soon. OpenAI also confirmed the changes in ChatGPT release notes on 10 February and announced a more minor GPT-5.2 update focused on more measured responses.

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

Threads users can now personalise algorithms with Dear Algo feature

Meta has launched an AI feature called Dear Algo that allows Threads users to personalise their content-recommendation algorithms by communicating their preferences directly to the platform.

Users craft public posts beginning with ‘Dear Algo’ to explain desired content, similar to interactions with chatbots like OpenAI’s ChatGPT. Once shared, the request adjusts the user’s feed for three days, allowing them to stay connected to current conversations.

Users can also repost others’ Dear Algo requests to apply those content preferences to their own feeds.

The feature represents Meta’s continued integration of AI across its platforms. The company on Tuesday also released AI features for Facebook that let users animate profile photos and alter images with the Meta AI digital assistant.

Meta told investors last month it plans to spend between $115 billion and $135 billion in 2026 on AI-related capital expenditures, nearly double last year’s spending.

Meta is testing Dear Algo with users in the US, UK, Australia, and New Zealand before expanding to additional countries. Threads, Meta’s micro-blogging platform launched in July 2023, has 400 million monthly active users and began rolling out advertisements globally last month.

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

AWS chief sees AI shifting from content creation to autonomous task completion

AI is shifting from answering questions to autonomously accomplishing tasks, a transformation AWS CEO Matt Garman believes will unlock far greater enterprise value.

Speaking at AWS re:Invent 2025, Garman explained that AI inference- the computing capability that allows models to generate content, make predictions, and take actions against real-world data-represents a fundamental new building block in computing.

He described it as developers gaining access to a ‘new Lego’ that enables applications to make decisions and complete work independently. The distinction between content generation and task accomplishment carries significant implications for enterprise value.

First-wave generative AI focused on writing emails and summarising documents. Task-accomplishing agents can review insurance claims, cross-reference medical records, and process approved claims without human intervention.

Garman predicts widespread enterprise value creation from agents in 2026. AWS announced Amazon Bedrock AgentCore and three frontier agents at re:Invent 2025, providing organisations with infrastructure to deploy autonomous AI agents at scale.

For business leaders, investments in agents that automate end-to-end workflows will deliver exponentially more return on investment than tools that help employees work faster.

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

Women driving tech innovation as Web Summit marks 10 years

Web Summit’s Women in Tech programme marked a decade of work in Qatar by highlighting steady progress in female participation across global technology sectors.

The Web Summit event recorded an increase in women-founded startups and reflected rising engagement in Qatar, where female founders reached 38 percent.

Leaders from the initiative noted how supportive networks, mentorship, and access to role models are reshaping opportunities for women in technology and entrepreneurship.

Speakers from IBM and other companies focused on the importance of AI skills in shaping the future workforce. They argued that adequate preparation depends on understanding how AI shapes everyday roles, rather than relying solely on technical tools.

IBM’s SkillsBuild platform continues to partner with universities, schools, and nonprofit groups to expand access to recognised AI credentials that can support higher earning potential and new career pathways.

Another feature of the event was its emphasis on inclusion as a driver of innovation. The African Women in Technology initiative, led by Anie Akpe, is working to offer free training in cybersecurity and AI so women in emerging markets can benefit from new digital opportunities.

These efforts aim to support business growth at every level, even for women operating in local markets, who can use technology to reach wider communities.

Female founders also used the platform to showcase new health technology solutions.

ScreenMe, a Qatari company founded by Dr Golnoush Golsharazi, presented its reproductive microbiome testing service, created in response to long-standing gaps in women’s health research and screening.

Organisers expressed confidence that women-led innovation will expand across the region, supported by rising investment and continuing visibility at major global events.

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

AI playlist creator comes to Youtube for Premium subscribers

YouTube has introduced a new AI Playlist feature for YouTube Premium and YouTube Music Premium subscribers on Android and iOS, enabling users to generate customised music playlists by describing a mood, genre, activity or vibe in natural language.

From the Library tab, users can tap ‘New,’ select ‘AI playlist’, and enter text or voice prompts, such as ‘sad post-rock’ or ’90s classic hits,’ to instantly build a curated list of tracks.

The rollout builds on YouTube’s earlier AI experiments in music discovery and positions the company alongside other streaming services like Spotify, Amazon Music and Deezer, which have launched similar generative playlist tools.

The feature reflects a broader trend of streaming platforms embedding generative AI to personalise discovery and enhance user engagement for paying subscribers.

Details such as the degree of user control over generated playlists and support for iterative refinement remain limited, and YouTube has not clarified how often playlists can be refreshed or edited after creation.

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

Global coalition demands ban on AI-nudification tools over child-safety fears

More than 100 organisations have urged governments to outlaw AI-nudification tools after a surge in non-consensual digital images.

Groups such as Amnesty International, the European Commission, and Interpol argue that the technology now fuels harmful practices that undermine human dignity and child safety. Their concerns intensified after the Grok nudification scandal, where users created sexualised images from ordinary photographs.

Campaigners warn that the tools often target women and children instead of staying within any claimed adult-only environment. Millions of manipulated images have circulated across social platforms, with many linked to blackmail, coercion and child sexual abuse material.

Experts say the trauma caused by these AI images is no less serious because the abuse occurs online.

Organisations within the coalition maintain that tech companies already possess the ability to detect and block such material but have failed to apply essential safeguards.

They want developers and platforms to be held accountable and believe that strict prohibitions are now necessary to prevent further exploitation. Advocates argue that meaningful action is overdue and that protection of users must take precedence over commercial interests.

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

AI controls animal behaviour using light-guided technology

Scientists at Nagoya University have developed an advanced AI system capable of identifying specific animal behaviours with over 90% accuracy and controlling the brain circuits that drive them in real-time across multiple species.

The system, named YORU (Your Optimal Recognition Utility), recognises entire behaviours from single video frames rather than tracking individual body parts, making it 30% faster than previous tools.

Researchers demonstrated the technology’s precision by combining it with optogenetics to silence a male fruit fly’s courtship song mid-performance, causing the unimpressed female to walk away.

The breakthrough lies in the system’s ability to target individual animals within social groups, so previous optogenetic methods illuminated entire laboratory chambers, affecting all subjects simultaneously.

YORU’s AI-driven light source can now track and manipulate a single subject’s neurons whilst its neighbours move freely nearby. The tool has proven its versatility across diverse species, successfully analysing food-sharing in ants, social orientation in zebrafish, and grooming patterns in mice.

Requiring minimal training data and no programming skills, YORU is available online for researchers worldwide studying the neural mechanisms underlying social interactions.

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

EU launches cyberbullying action plan to protect children online

The European Commission has launched an Action Plan Against Cyberbullying aimed at protecting the mental health and well-being of children and teenagers online across the EU. The initiative focuses on reporting access, national coordination, and prevention.

A central element is the development of an EU-wide reporting app that would allow victims to report cyberbullying, receive support, and safely store evidence. The Commission will provide a blueprint for Member States to adapt and link to national helplines.

To ensure consistent protection, Member States are encouraged to adopt a shared understanding of cyberbullying and develop national action plans. This would support comparable data collection and a more coordinated EU response.

The Action Plan builds on existing legislation, including the Digital Services Act, the Audiovisual Media Services Directive, and the AI Act. Updated guidelines will strengthen platform obligations and address AI-enabled forms of abuse.

Prevention and education are also prioritised through expanded resources for schools and families via Safer Internet Centres and the Better Internet for Kids platform. The Commission will implement the plan with Member States, industry, civil society, and children.

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

Microsoft explores superconductors for AI data centres

Microsoft is studying high-temperature superconductors to transmit electricity to its AI data centres in the US. The company says zero-resistance cables could reduce power losses and eliminate heat generated during transmission.

High-temperature superconductors can carry large currents through compact cables, potentially cutting space requirements for substations and overhead lines. Microsoft argues that denser infrastructure could support expanding AI workloads across the US.

The main obstacle is cooling, as superconducting materials must operate at extremely low temperatures using cryogenic systems. Even high-temperature variants require conditions near minus 200 degrees Celsius.

Rising electricity demand from AI systems has strained grids in the US, prompting political scrutiny and industry pledges to fund infrastructure upgrades. Microsoft says efficiency gains could ease pressure while it develops additional power solutions.

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