CUTS Daily Bulletin # 01 | July 08, 2025
Delegated Decisions, Amplified Risks: Charting a Secure Future for Agentic AI
 
Meredith Whittaker (President, Signal Foundation), Kenneth Cukier (Deputy Executive Editor, The Economist) led a critical conversation on the rising risks of agentic AI. Whittaker examined the security, privacy, and structural concerns of delegating decisions to AI agents. Cukier moderated the session with probing questions around implementation, governance, and real-world harm.
  1. Deep Access Across Systems: Agentic AI requires broad permissions across calendars, contacts, and apps—raising systemic privacy and security risks.
  2. Undermining Application Layer Protections: AI agents may breach secure environments like Signal, threatening trusted digital communication platforms.
  3. Data Exploitation Risks: AI intermediaries open new doors to surveillance, corporate overreach, and competitive data misuse.
  4. Prompt Injection Vulnerabilities: Agents can be tricked by malicious prompts, creating unintended and unsafe actions.
  5. Implementation Over Principle: The problem is not AI itself, but how it is built—often prioritizing hype and monetization over safety.
  6. Lack of Verified Components: Many AI systems are built on unverified libraries, exposing core vulnerabilities.
  7. Urgent Policy Solutions Needed: Whittaker called for developer opt-outs, formal verification, and procurement standards to safeguard infrastructure.
  8. Ethical Imperative: Without stronger regulation and public awareness, privacy and autonomy could become irretrievable losses in the age of AI.
(Reporting by Peter Maundu, CUTS International, Geneva)
AI for Good – Food and Agriculture
 
Dr. Qu Dongyu, the Director-General, Food and Agriculture Organization (FAO), addressed the urgent need to bridge the global digital divide, particularly in rural and low-income regions. At the AI for Good Summit 2025, he presented FAO’s vision for ethical, inclusive, and effective AI in transforming agro-food systems.
  1. Digital Exclusion Threatens Development: Nearly 2.6 billion people remain offline, mostly in rural and low-income countries, limiting their participation in the AI revolution.
  2. AI for Sustainable Farming: FAO promotes AI to help smallholder farmers increase yields with fewer resources, while enhancing climate adaptation and resource management.
  3. Geospatial Tools in Action: AI-driven remote sensing allows FAO to monitor droughts, water scarcity, and land use efficiently.
  4. Localized AI Advisory Services: Through partnerships like Digital Green, FAO delivers cost-effective, AI-powered crop advice in local languages.
  5. Open Knowledge Access: FAO’s open-access policy and AI knowledge box democratize data for researchers, policymakers, and farmers.
  6. Youth Innovation Empowered: The Robotics for Good Youth Challenge invites youth to develop AI-driven solutions for food security, supported by FAO’s mentorship.
  7. Ethics at the Core: FAO emphasizes AI governance rooted in human dignity, transparency, and the principle of “do no harm,” as part of its ethical AI agenda.

(Reporting by Peter Maundu, CUTS International, Geneva)
Optimism for AI – Leading with Empathy
 
Will.i.am (Musician, Founder & CEO FYI.AI) and Nicholas Thompson (CEO, The Atlantic) engaged in a thought-provoking conversation about the intersection of creativity, ethics, and AI. Will.i.am emphasized the need to lead with empathy and human intuition in AI development, warning against cognitive offloading and creativity outsourcing. The discussion combined deep technical insights with cultural reflections on AI’s future direction.
  1. Human Creativity Comes First: AI should amplify, not replace, human imagination. Will.i.am uses AI to generate variations, but insists the original idea must come from people.
  2. Pattern Matching as a Superpower: Creativity and innovation come from recognizing patterns. Will.i.am sees pattern-matching as a transferable skill across music, visuals, and technology.
  3. AI Constitution and Licensing: A regulatory framework is essential. Will.i.am proposed an “AI Constitution” and licensing for developers, like driving permits, to ensure public safety.
  4. Ethics vs. Greed: Many AI systems are being developed with profit as the primary motive. Will.i.am calls for human-first innovation led by ethical responsibility.
  5. Empowering Youth: Education must foster ambition, audacity, and competitive spirit. Will.i.am’s AI education initiatives prepare underrepresented youth for the future of work.
  6. Precaution with Power: AI is powerful and transformative—but without empathetic leadership, it risks becoming dangerously inhumane.

(Reporting by Peter Maundu, CUTS International, Geneva)
Early Warning for All – Leveraging AI to Identify the Unconnected
 
This session presented a groundbreaking collaboration between ITU, Microsoft, IHME, and Planet Labs to map unconnected populations using AI and satellite data. Doreen Bogdan-Martin highlighted the urgency of ensuring early warning systems reach every person by 2027, in line with the UN’s “Early Warnings for All” initiative.
  1. Mapping the Unconnected: AI and satellite imagery now allow precise identification of populations beyond mobile coverage, enabling more inclusive early warning systems.
  2. From Vision to Tool: The Early Warning Connectivity Map (ECM) combines connectivity data and high-resolution population maps to reveal underserved areas.
  3. AI-Driven Risk Modeling: IHME and Microsoft use demographic models and AI to anticipate threats like climate impacts, health risks, and migration.
  4. Daily Satellite Insights: Planet Labs provides real-time Earth imagery, feeding AI systems with continuous, global population intelligence.
  5. Precision Collaboration Programme: A new early-access program invites partners to leverage these insights to build resilient emergency systems.
  6. Data Equity as Lifesaving Infrastructure: Speakers emphasized that open, shared, and localized data can radically improve global preparedness, equity, and response.
  7. Call to Action: The session closed with an invitation to join the AI for Early Warning for All group to scale global solutions.
The session featured speakers Doreen Bogdan-Martin (ITU), Juan Lavista Ferres (Microsoft), Prof. Christopher Murray (IHME), and Andrew Zolli (Planet Labs), and was moderated by Vanessa Gray (ITU).

(Reporting by Peter Maundu, CUTS International, Geneva)
Unmapped: No Data, No AI
 
This report summarizes the session "Unmapped: No Data, No AI" presented by Werner Vogels, Chief Technology Officer and Vice President of Amazon, at the AI for Good Summit 2025. Dr. Werner Vogels underscored the critical role of maps as foundational data for AI to function meaningfully in the real world. Through compelling global examples, he demonstrated how open, shared geospatial data drives innovation in health, climate, and disaster resilience. His closing message: transparency and collaboration are essential—ask not only what data you need, but what data you can share. The Key takeaways were as follows:
  1. Empowering Local Communities Through Data and Technology: Apps like Grab and Yamamatra empower communities to map underserved areas using mobile and IoT tools. These citizen-led efforts fill data gaps, improving logistics and civic infrastructure.
  2. Fellowship Spotlight: Now Go Build CTOs: This fellowship supports technologists applying AI to critical issues like agriculture and disaster response, fostering scalable, community-led solutions.
  3. Rwanda’s Real-Time Health Mapping: Rwanda uses mapping data to ensure equitable health access, including planning clinics within an hour’s walk for pregnant women.
  4. Ocean Cleanup: Predictive Plastic Tracking: Using AI and drone data, the initiative models plastic flow in rivers to optimize cleanup operations globally.
  5. Cloud Infrastructure at Global Scale: AWS enables analysis of petabyte-scale data and offers free access to essential datasets through its Open Data Sponsorship Programme.
  6. CLAI: Earth Observation with Open AI: CLAI uses open geospatial data to track environmental changes, proving the power of accessible, responsible AI.
  7. The Ethics and Urgency of Open Data: Dr. Vogels called data hoarding unethical, urging global data sharing to tackle shared challenges.
  8. Geospatial Data: The Hidden Engine of the SDGs: All 17 SDGs depend on mapping. Without geospatial data, progress cannot be measured—or achieved.

(Reporting by Peter Maundu, CUTS International, Geneva)
Take your first step towards change with CUTS International
CUTS International

Jaipur • New Delhi • Chittorgarh • Kolkata • Hanoi • Nairobi • Lusaka • Accra • Geneva • Washington DC

Copyright © 2025 CUTS International, All rights reserved.