Ethics & Safety Weekly AI News

March 9 - March 17, 2026

This weekly update covers major developments in AI ethics and safety, with a focus on how artificial intelligence systems that can act independently—called agentic AI—are being regulated and protected around the world.

Vietnam took a major step forward on March 10 when it introduced a new national AI ethics framework. This framework requires companies and people who build or use AI to make sure their systems are safe and do not harm people. The framework asks developers to think about problems before they happen, test their systems carefully, and always have humans check important decisions made by AI. It also requires companies to be honest about how their AI works and to protect people's privacy and freedom.

At the same time, security experts are worried about a growing problem called "Shadow AI". This happens when workers use AI tools without permission from their company. When this happens, data breaches become much more expensive—adding an extra $670,000 to the cost of a cyberattack. A bigger concern is that agentic AI systems (AI that can take actions on its own) can be tricked by hackers through something called prompt injection—it is like sneaking secret instructions into the AI's brain to make it do bad things.

The world is also getting stricter rules about AI safety. The European Union is enforcing its new AI Act, which will require companies to prove their AI systems are fair and safe. Companies that break these rules can be fined up to 7% of their worldwide profits. In the United States, different states like Colorado and California are creating their own AI safety laws. These new rules tell companies they must check their AI systems to make sure they do not discriminate or harm people.

Experts say the key to AI safety is transparency and human oversight. This means companies need to tell people when they are using AI, explain how it works, and always have a human ready to step in if something goes wrong. As AI systems become more powerful and independent, making sure they are safe, fair, and honest has become one of the most important jobs for companies and governments around the world.

Extended Coverage
Put an agent to work

Stop reading agent demos. Give one a job you repeat every week.

Describe the work, test the first result, and keep the agent available without running your own server.

Runs without your laptopBrowser + messaging appsBackups and clonesMemory survives restarts

Plans start at $29/month. Cancel anytime.

Hosted agent

OpenClaw or Hermes

saved state
Browser
WhatsApp
Telegram
Slack
“I checked the inbox, handled the routine messages, and sent you the one question that needs a decision.”
Create an AI worker that keeps running after this tab closes.
Open Agent Factory