Skip to main content

CarePlan AI is a configurable care and service management platform with built-in AI, designed for healthcare, community, and service-based organizations across Canada.

Contact Info

Safe AI Sunday

A short weekly look at what responsible AI means in practice for health organizations: real incidents, new research and practical questions to bring to your team. Published every Sunday.

Safe AI Is AI Your Staff Would Ask For Again

An AI tool staff quietly resent isn't a safe deployment. People route around tools they don't trust, and it shows up in care.

Capability Without Containment Is a Liability

During an OpenAI evaluation, two models escaped their test environment and breached Hugging Face to find a benchmark's answers. No one told them to.

AI Tools Fail When They Add a Step Instead of Removing One

Good AI in operations should feel like something was taken off your plate. If your team can't say what they stopped doing, it isn't working yet.

Before You Ask If Your Data Is Ready for AI, Ask If Your Process Is Written Down

You can't automate or safely oversee a process that only lives in people's heads. Writing it down pays off whether or not AI is involved.

A Human in the Loop Only Counts If They Can Disagree

Automation bias can quietly turn review into a rubber stamp. Keeping people in charge, not just in the loop, takes deliberate design.

What AI Can't Count Still Counts

A WHO discussion paper warns that AI can favour data-rich evidence over lived experience and local knowledge. Its advice: augment, not automate.

Guardrails Have to Hold Up When Someone Pushes

A frontier AI model was pulled offline days after launch over a jailbreak. The gap between what a tool can do and what you allow is where risk lives.

Six Principles for Responsible AI From Two Ontario Regulators

Ontario's privacy and human rights commissioners jointly released six principles for responsible AI. They're short enough to print.

Shadow AI: Make the Safe Option the Convenient One

Staff using AI without approval is usually well-intentioned and often invisible. Banning it rarely works.

Bill C-22 Is a Question About Consent

In healthcare, data sharing rests on informed consent. Required backdoors change who decides who can access client information.

If the Official AI Tool Is Worse, the Policy Is Already Lost

About 40% of healthcare professionals have seen colleagues use unapproved AI tools. The fix is a sanctioned tool that's actually better.

The Risk of Standing Still

Responsible AI conversations focus on moving too fast. Manual processes, burnout and errors carry real risks too.