|
|
|
HIGHLIGHTS
Citigroup pilots AI agents for employees. Citi has launched a pilot with 5,000 staff using “agentic” AI integrated into its in-house platform. These agents can execute multi-step tasks from a single prompt, such as gathering client data, analyzing it, and producing multilingual reports. The bank is testing productivity gains against strict cost controls before scaling adoption, acknowledging the potential for future workforce impact. WSJ
|
Nvidia powers global “sovereign AI” ambitions. CEO Jensen Huang is securing national deals to build AI supercomputing hubs, pledging £2 billion to UK startups and investing £500 million in London’s Nscale facility. Over 20 countries have signed sovereign cloud agreements this year, making Nvidia indispensable to governments seeking AI independence. While marketed as sovereignty, the reality is growing reliance on one U.S. company’s chips. FT
|
Factory automation in overdrive in China. Chinese manufacturers are pushing “lights-out” factories, with executives declaring that entire production lines may soon operate without a single human worker. Supported by state incentives, companies are rapidly adopting robots and AI systems to cut labor costs and increase export competitiveness. Human staff are being retrained, but the trend points to AI-driven dominance in global manufacturing. FT
|
OTHER NEWS
-
|
How people are using GPTs. OpenAI’s analysis shows users rely most on GPTs for productivity tasks like writing, coding, and tutoring, with creative exploration and personal projects also gaining traction. OpenAI
|
-
|
Is AI hype outpacing reality? An analysis warns that AI’s economic promise may be exaggerated, citing failed pilots and slowing productivity gains, though some experts point to a “J-curve” delay before benefits emerge. The Atlantic
|
-
|
How to make enterprise GenAI work. A new Harvard Business Review article highlights why most pilots fail, pointing to weak problem framing, poor change management, and lack of integration into workflows, and offers a roadmap for scaling success. HBR
|
|
|
HIGHLIGHTS
AI-designed life forms. Researchers used generative AI to design whole genomes for new viruses (bacteriophages) and brought them to life in the lab. Out of 300 AI-generated designs, 16 created viable viruses, some of which killed bacteria faster than natural ones. The breakthrough could transform phage therapy but raises urgent biosafety questions. Article
|
Teaching AI personalities. A University of Pennsylvania study gave AI agents personality traits like introversion or extroversion using the Myers-Briggs framework. Introverted AIs produced more reflective responses, while extroverted ones were bolder and more spontaneous, and diverse teams of “personalities” worked better together. The findings suggest tailoring AI “psychology” could align systems more closely with human values and roles. Article.
|
SML as a new path for agentic AI. Researchers propose Small Language Models (SMLs), lightweight models with a few hundred million parameters, as efficient controllers for agentic AI. Rather than doing everything themselves, SMLs orchestrate specialized tools and APIs, enabling modular, auditable, and cost-effective reasoning. The approach promises more controllable agents, bridging the gap between today’s large LLMs and practical enterprise-scale systems. Article
|
OTHER NEWS
- Graph Neural Networks explained. Unite.AI highlights how GNNs are enabling breakthroughs in fraud detection, drug discovery, and traffic optimization by modeling relationships in complex networks. Unite.AI
- K2-Think: new benchmark for reasoning. A new research initiative introduces K2-Think, a large-scale benchmark designed to measure multi-step reasoning across diverse tasks, aiming to push LLMs closer to true problem-solving intelligence. Article
- Global AI policy — a third way. A Brookings report proposes a “third AI stack” where democracies pool resources to counterbalance U.S. and Chinese ecosystems with shared governance and transparency. Brookings
|
|
HIGHLIGHTS
NoeSysAI hosts workshop with CDO and Data Governance leaders. This month we brought together Chief Data Officers and Data Governance professionals for an in-depth workshop on AI strategy and governance. The session focused on practical challenges of enabling responsible AI adoption, from data readiness to organizational change, and gave participants hands-on frameworks to shape their company’s AI journey. LinkedIn
|
The Age of Intelligence podcast — UAE’s AI vision and rethinking the “Empire of AI.” Theodoros Evgeniou co-hosted two powerful episodes of The Age of Intelligence. With Saif AlSalman, Microsoft’s National Technology Officer for the UAE, the discussion explored how the country is making AI a national priority, from $1.4 trillion in investment to building Arabic LLMs. With journalist Karen Hao, author of Empire of AI, the focus turned to power structures in AI, the risks of hype and hidden labor, and why shaping a more equitable AI future matters. Podcast with Saif AlSalman - Podcast with Karen Hao
|
How AI is transforming HR. In an article for INSEAD Knowledge, Theodoros Evgeniou, Gaël Gioux and Elie Bechara (LinkedIn) highlight how AI is reshaping the HR function. From talent acquisition to workforce planning and employee engagement, HR leaders are increasingly on the front line of AI-driven transformation, requiring new skills, governance frameworks, and ethical guardrails. INSEAD Knowledge
|
OTHER NEWS
- Our C-Level AI Value Creation Workshop. We have officially launched our customized C-Level AI Value Creation Workshop, helping executive teams rapidly identify and capture AI opportunities tailored to their industries. Link
|
|
|
|
|
|
