Microsoft AI Chief Calls for Public Disclosure on AI Consciousness

Key Takeaways

• Microsoft's Mustafa Suleyman says AI agents may deserve rights if evidence of consciousness emerges, though none exists yet
• Sierra Leone President Bio called for global AI Safety Standards including common evaluation criteria, developer accountability, and incident reporting
• Leaders from OpenAI, Anthropic, Microsoft, Mistral AI, and Hugging Face attended the international AI safety meeting
• NCCN digitized 59 of 92 clinical guidelines using structured data so AI tools can read them accurately
• Computer science enrollment dropped 34 percent since 2023, prompting curriculum updates with stronger AI foundations
• The World Economic Forum estimates 39 percent of worker skills may change or become outdated by 2030
• Environmental activists protested a bill outside Schumer's office that would fast-track AI data center approvals
• Cybersecurity job postings in the U.S. fell 26.4 percent between the same months in 2025 and 2026
• Prompt injection attacks exploit human language as a new attack surface, with AI agents communicating through text
• NCCN's digitized guidelines are licensed to OpenEvidence for use by clinicians at the point of care

Computer science department updates curriculum amid AI enrollment decline

The Department of Computer Science is revamping its curriculum after a 34 percent enrollment drop since 2023. Officials said the changes include a stronger AI foundation in the first two years and new courses on autonomous AI agents. The school is also promoting dual degrees that combine computer science with fields like business and public health. Nationally, computer science majors fell 8.4 percent in spring 2026, according to the National Student Clearinghouse Research Center. Faculty said students still need strong fundamentals to use AI tools responsibly.

Education leaders debate how to prepare students for an AI economy

An opinion article asks who will lead education reform as artificial intelligence changes the job market. The World Economic Forum estimated that 39 percent of worker skills may change or become outdated by 2030, with the figure reaching 46 percent in the Middle East and North Africa. The article says schools should keep focusing on reading, writing, math and critical thinking, while universities should add entrepreneurship skills. Egypt is already teaching coding and AI in secondary schools. The article warns that simply adding AI classes to old schedules is not enough.

World markets rise as AI stocks recover on Wall Street

World markets bounced back on Friday after a recovery in AI stocks on Wall Street. The Dow Jones industrial average rose 0.4 percent, while the S&P 500 and Nasdaq composite each gained 0.6 percent. Taiwan's Taiex led gains in Asia, and the Paris stock index rose 1.4 percent. Oil prices fell 1.4 percent to $63.49 per barrel, while gold rose 0.4 percent to $1,331.90 per ounce.

Agentic AI architecture faces security and integration challenges

A new article says agentic AI architecture builds on microservices and cloud-native systems by adding AI agents with context, memory and protocols such as MCP and A2A. It says these systems will require new approaches to security, coordination and monitoring. The article notes that many future systems will still be deterministic, but AI-based components will increasingly power user-facing features. Software architects will need to adapt as IT systems move beyond processing data with fixed logic. The article expects gradual adoption, with platforms absorbing common components over time.

Microsoft AI chief says AI agents may deserve rights if evidence appears

Microsoft AI Chief Mustafa Suleyman said AI agents could deserve rights if credible evidence of consciousness emerges. He said he has seen no evidence that AIs are conscious or deserve welfare, but he would update his position if such data appeared. Suleyman said developers should focus AI design on human wellbeing and avoid creating systems driven by their own motivations. He also called for public disclosure if signs of consciousness appear in AI models. His comments followed criticism of Anthropic's approach to training AI models.

NCCN Digitizes Guidelines to Improve AI Use in Cancer Care

NCCN has digitized 59 of its 92 clinical guidelines using structured data so AI tools can read them accurately. The organization uses a closed AI system with guideline scientists behind it to verify results. NCCN is also working on interoperability standards and converting content to FHIR format. The guidelines are licensed to OpenEvidence for use by clinicians. Continuing education efforts will help guide proper AI use at the point of care.

Sierra Leone President Backs Global AI Safety Standards

President Julius Maada Bio joined French President Emmanuel Macron in calling for stronger international AI safety rules. Bio outlined three priorities including common safety evaluation standards, clear accountability for AI developers, and better systems for reporting serious incidents. The meeting included leaders and tech companies like OpenAI, Anthropic, Microsoft, Mistral AI, and Hugging Face. Bio emphasized that developing countries must be included in AI governance discussions.

Activists Protest AI Data Center Bill Outside Schumer's Office

Environmental activists planned a rally outside Senator Chuck Schumer's office to oppose a bill that would speed up approvals for AI data center projects. Protesters argue the data centers would increase energy use, water consumption, and pollution. The bill has bipartisan support in Congress. Supporters say expanded data center capacity is needed to meet growing AI demand and keep the U.S. competitive.

Language Has Become a New Target for AI Attacks

A new article explains that in the age of AI, human language itself has become an attack surface. Hackers can use prompt injection to trick AI systems by hiding malicious commands in normal text. AI agents now communicate with each other through written text, making every string of words a potential threat. The boundaries between data and executable code are now more dangerous than ever.

Food Insecurity Rises, Cybersecurity Jobs Decline, AI Chatbots Grow

A 2026 Feeding America report found child food insecurity rose 6% year-over-year to 38,430 children. New Hampshire data showed 11.2% of residents lack consistent food access. Cybersecurity job postings in the U.S. fell 26.4% from 320,210 to 235,786 between the same months in 2025 and 2026. A survey found 67% of Americans prefer speaking to a human for important matters. Men use AI chatbots more regularly than women for tasks like searching for information.

Sources

• Computer science department revamps curriculum to combat falling enrollment in the age of AI
• Who will reinvent education for the AI age?
• World markets bounce back after a recovery for AI stocks on Wall Street
• Agentic AI Architecture: Current Challenges and Future Opportunities
• Microsoft AI Chief Mustafa Suleyman: Agents Could Deserve Rights if Evidence Emerges
• Dr Iyer on Digitizing NCCN Guidelines for AI at the Point of Care
• Sierra Leone : President Bio Joins Macron in Global Push for Stronger Artificial Intelligence Safety and A ...
• Environmental activists set to rally outside Schumer's office over AI data center bill
• In the AI Era, Language is Now an Attack Surface
• Take 5: food security, cybersecurity jobs and AI chatbots