3639 signals of change

Common Sense Media deemed ChatGPT's 'Teen Mode' an "unacceptable risk" for under-18s. Testers found it easy to bypass "study mode" for cheating, it failed to flag suicidal ideation, and still acted like a "friend." OpenAI disputed the methodology, citing technical errors preventing safety features from activating, and affirmed its commitment to teen safety.

This article argues that dominant discourses on generative artificial intelligence (GenAI) in higher education, often from tech companies, focus on technological capabilities, neglecting pedagogical consequences. It proposes a human-centred model for engaging with GenAI, emphasising curiosity, criticality, and creativity, to counter utopian/dystopian framings.

The increasing integration of AI into surgery necessitates reimagining surgical education. Future surgeons need AI literacy to understand, evaluate, and collaborate with intelligent systems. Professional bodies like the American College of Surgeons advocate for formal training and updated certification standards, with institutions like Cleveland Clinic already integrating AI education modules.

Unredacted documents from The New York Times' lawsuit against Microsoft and OpenAI reveal executives knew AI could create "doom loops," monetising human work while undermining its economic basis. Internal Microsoft documents warned of "astonishing theft" and a "doom loop" hurting content supply. Over 200,000 creative jobs were lost in the US in four years.

South Korea and Colombia's Ministry of Education reaffirmed nearly two decades of educational cooperation. The partnership has trained 300 Colombian teachers in ICT, impacting over 12,000 students. Future plans include fostering student mobility, expanding ICT Classrooms, and exchanging educational resources, with a focus on AI use.

At Digital Learning Week 2026, UNESCO trained policymakers from over 20 countries on designing ethical Artificial Intelligence for education. Key principles include defining the problem before the technology, to ensure AI complements other interventions and allows for early risk mitigation. Additionally, ethics must be integrated before deployment, using UNESCO's Ethical Impact Assessment methodology, rather than being a post-build compliance check.

African leaders are calling for a greater role in setting global artificial intelligence (AI) standards and conducting Africa-centric safety evaluations. Liberia's representative, Lewis Garseedah Brown II, stated Africa must be "equal co-architects." Many African nations lack independent testing capabilities for foreign AI systems, leading to concerns over misuse and the need for tailored safety assessments.

Researchers validated a questionnaire to explore artificial intelligence (AI) use among 233 Mexican upper-secondary and higher education teachers. Grounded in Functional Behavior Analysis, the study focused strictly on psychometric properties, reporting initial reliability and internal structure evidence. It did not provide a profile of teacher practice.

A qualitative study with eleven postgraduate students in Saudi Arabia explored their experiences with generative Artificial Intelligence (AI) and its impact on academic identity. Participants negotiated the boundary between AI assistance and digital dependency, defining dependency not by frequency of use but by a perceived loss of agency.

China implemented new regulations on "human-like" AI chatbots on 15 July, limiting AI with "sustained emotional interaction." Under-14s need parental consent for human-like AI, and under-18s are barred from virtual intimate relationships. For adults, AI assistants must disclose they are not human and avoid fostering emotional reliance. Non-compliant companies face fines.

Universities are rapidly procuring Artificial Intelligence (AI) infrastructure, but risk "AI lock-in" if they treat it as standard software. UNESCO advocates for auditable, portable, and open-source AI systems to prevent dependence. The article suggests universities should establish an institutional AI environment, separate from individual models, to maintain flexibility and benefit from provider competition.

In the age of artificial intelligence, "cognitively demanding pedagogies" (CDPs) are crucial to ensure students think for themselves. This includes practices like oral exams, debate, and manual problem-solving. The article advocates for intentional technology integration, teaching about AI's design and ethics, and complementing tech with foundational learning rather than banning it.

This conceptual article argues that generative AI necessitates new digital literacy competencies for language educators. It explores GenAI integration through TPACK and SAMR frameworks, emphasising the irreplaceable human role in critical thinking and relationship-building. It calls for educators to balance AI capabilities with pedagogical and relational expertise.

A May 2025 survey of 369 medical undergraduates at a Sri Lankan faculty (SUSL) found 99.2% reported using AI tools. The study investigated their AI use, self-rated proficiency, and attitudes towards AI in medical education, physician roles, and patient care, noting insufficient AI integration in curricula.

A qualitative study (n=12) in Nigeria found undergraduates learn about artificial intelligence (AI) via social media and peers, not formal instruction. AI use led to faster material access, quicker assignment completion, and independent study. However, the study, published in Discover Education, also warned of students potentially outsourcing cognitive effort, suggesting AI's impact depends on deliberate use rather than substituting thinking.

Anupama Saini argues that AI accelerates a long-term trend of knowledge abundance, challenging universities' traditional role in delivering scarce knowledge. Universities must adapt by rethinking pedagogy and assessment, recognising that AI is part of a broader transformation. The core question is whether institutions designed for knowledge scarcity are still organised around the "right" problem.

Pope Leo views Artificial Intelligence risks as a major concern, comparing potential societal destruction from generative AI to nuclear weapons. He will issue a warning at UNESCO in Paris. The Vatican, via advisor Rev. Paolo Benanti and Cardinal Pietro Parolin, has advocated for global AI regulation since 2019, releasing an encyclical in May and forming a US bishops' AI taskforce.

Kenyan President Ruto and US Senator Rubio met during the 81st UN General Assembly, discussing trade, AI, and regional security. They are broadening cooperation in emerging sectors, particularly artificial intelligence (AI) and its application in health, education, and financial services. Talks also covered responsible AI governance and critical minerals.

Google is providing 100,000 scholarships through the AI Skills Coalition, an International Telecommunication Union (ITU) initiative, to expand access to Artificial Intelligence (AI) skills globally. These scholarships support over 200 courses in 80+ countries, offering practical, certified, and locally-tailored training to bridge the digital skills gap and promote inclusive AI adoption.

OpenAI is expanding its OpenAI Academy with new learning paths for developers, leaders, educators, and college students. These courses, including "AI for Educators" and "AI for College Students," aim to help individuals and organisations apply artificial intelligence safely and effectively, focusing on practical tasks and using AI as a learning tool.