👋🏻 Hello! 2,090+ readers use ”Sensible AI” to move from headlines to decisions.
Starting today you’ll get 2 shorter editions (Tuesday and Friday).
In this 9th edition, you’ll see signals of global governance, companies tightening security and the Reddit vs. Perplexity case reopening the data-and-licensing debate.
As always, you’ll get “from the news to your action plan” the risk of the week and a key resource.
Let’s begin 👇
TOP NEWS
🇨🇳 China creates WAICO for global AI governance
→ Announces the World Artificial Intelligence Cooperation Organisation to coordinate international standards.
🇲🇽 Mexico launches PotencIA MX
→ Meta, Tec de Monterrey and the government create the first national open-AI accelerator to boost 30 startups with no licensing fees.
⚖️ Reddit sues Perplexity
→ Accuses the startup of training its AI on unlicensed content and seeks monetary damages.
🧠 Anthropic raises safety standards
→ Publishes its AI Scaling Policy with new auditing and simulation techniques to scale AI safely.
WHAT HAPPENED WORLDWIDE?
🏥 WHO & South Korea push safe AI in health
→ At AIRIS 2025, regulators and countries agree on a global strategy for ethical, equitable AI in healthcare.
🍁 Canada risks health data sovereignty
→ Delays to the AI Data Act leave sensitive and Indigenous health data exposed to AI without legal safeguards.
🇨🇴 Colombia promotes critical AI literacy
→ UNESCO launches Minds Over AI during Global MIL Week to strengthen media literacy against algorithmic manipulation.
🇦🇪 Dubai unveils AI plan targeting 30 unicorns
→ Sheikh Hamdan introduces a platform to create 3,000 jobs and position Dubai as a global tech hub.
💰 Banking sector evaluates risks from “agentic” AI
→ Financial services consider new governance frameworks to mitigate risks from autonomous systems in operations.
☁️ World Meteorological Organization adopts AI
→ WMO applies AI/ML to forecasts and warnings without replacing human judgment, aiming for global coverage and climate safety.
🧩 Adigital & IDB Lab present IAméricas Mexico
→ Program to promote ethical, responsible AI, aiming to reach 350,000 women and 1,500 organizations by 2027.
🇯🇵 Japan publishes AI Basic Plan
→ Sets policies for safe AI that balance innovation and risk, aligned with international standards.
🇦🇪 Abu Dhabi to host Global AI Summit
→ Will convene government and business leaders to discuss responsible governance and inclusive access.
🇮🇸 Iceland approves a National AI Plan
→ Adopts 25 measures for innovation, education and ethical governance aligned with the EU.
🇻🇳 Vietnam accelerates AI with safety
→ Strengthens transparency and accountability while pushing its tech industry forward.
🔬 Spanish scientists demand AI transparency
→ COSCE calls for regulating AI without replacing human clinical judgment, prioritizing ethics, evidence and safety.
🇰🇪 Kenya advances a global resolution on ethical AI
→ Proposes at the UN a safe, sustainable AI that serves people and the planet.
AND AT ENTERPRISE LEVEL?
🏛️ PROA Comunicación launches ethical AI in Spain
→ With BeAI, it creates the first confidential AI model for corporate reputation management.
🌍 ETHIS – global platform for human governance in AI
→ Launched during SF Tech Week, it promotes direct human control and citizen participation in AI regulation.
🧑💻 Microsoft unveils human-centered Copilot
→ The new version adds 12 features to personalize AI and improve collaboration and workplace well-being.
ORGANIZATIONS & STANDARDS
🌐 Global AI governance from WEF
→ WEF highlights the AGILE 2025 Index, which measures AI regulatory capacity across 40 countries and reveals key gaps.
FROM THE NEWS TO YOUR ACTION PLAN
I turn these headlines into concrete steps you can start today to strengthen your AI governance. Pick 3 actions to drive impact in 30 days:
🟪 AI Committee → create a cross-functional committee with veto power and a biweekly agenda for risks, roadmap and metrics.
🟪 BYOAI policy & training → publish a “bring your own AI” guide and train all staff on allowed uses, privacy and traceability.
🟪 Ethics & human rights → add a rights-impact and non-discrimination assessment at project kickoff and closure.
🟪 Vendors & exit → include audit clauses, data/model portability and a tested exit plan in all contracts.
🟪 Ecosystem & public voice → participate in regulatory consultations and sign MOUs with academia and civil society for independent testing and validation.
RECOMMENDED READS
📄 I recommend this report that clearly summarizes the progress of global regulatory frameworks, with a special focus on transparency and risk categorization. It highlights the momentum of the EU AI Act and international cooperation.
✍🏼 My notes :
I’m taking away the importance of prioritizing transparency to build trust and accountability and adopting flexible, adaptive frameworks to respond quickly to change.
RISK OF THE WEEK
🔴 How to opt-out
Imagine an analyst pastes a contract excerpt into their favorite chat “just to get a quick summary.” Weeks later, the same clause patterns start “appearing” in responses to other users because the provider kept and used those prompts to train its model. There was no intentional leak, but there was unnecessary exposure: sensitive data in a system you don’t control, traces that are hard to erase and potential conflicts with NDAs or privacy rules.
Opting out of data retention cuts that thread: your searches don’t feed the model, you reduce legal and technical risk surface and you protect your team’s know-how.
👉🏼 What I’d do in your shoes:
🟩 Disable “use my data for training/retention” in Settings → Privacy for the provider you use.
🟩 Declare it in your BYOAI policy.
🟩 If a service offers no opt-out, treat it as public: no PII, secrets, code, or contracts.
Need help?
If you’re not sure how to do this, no problem: we’ll create a plan together.
RESPONSIBLE HUMOR
That’s all for today!
What topic would you like me to dig into on Friday?
If this is useful, forward it to your team and reply with your feedback (it helps a lot). See you Friday 💌
|
With gratitude and tea in hand ☕, Karine Mom, passionate about human-centered Responsible AI – AI Governance Consultant & WiAIG Mexico Leader |
 |
All sources have been verified and dated. If you see any mistake, reply and I’ll fix it on the web version.