Short answer: it means three things, not one. You chose the tool, you read and sign every note, and you can change it or switch it off. Reading the draft is only a third of it. In September, about 150 doctors at two Minnesota hospitals went on strike, and a say in how AI is used was one of their demands.

“Human in the loop” is the phrase vendors use to reassure clinicians. The clinicians living with it are less sure it means anything on its own.

Why isn't reviewing the note enough?

Because the person reviewing did not choose what they are reviewing. On September 14, about 150 physicians at Allina Health's Mercy Hospital in Coon Rapids and its Unity campus in Fridley, Minnesota, began a four-day strike. Alongside sick time and benefits, their union, Doctors Council SEIU, asked for a guaranteed role in determining how AI is used in patient care.

One commenter on the strike, writing on a health IT news site, put it in a sentence:

“There isn't really a human in the loop when AI is imposed”

as just another enterprise system, the commenter went on, without clinicians' authority over how it is selected, monitored and used.

What were the striking doctors worried about?

Who decides, and what the software is for. Dr. John Wust, an OB-GYN on the union's bargaining team, told KARE 11 that the AI suggestions doctors were seeing were about revenue, and that they did not want to be “told how to code visits and how to bill patients” on profitability alone. His line on where AI belongs:

“It should be used as a tool, not as a final decision-maker.”

Dr. Dawn Anderson, an internal medicine physician, named the other fear: that it would keep eroding their time at the bedside.

The tentative agreement reached on September 24 did not include specific terms on AI, MPR News reported. It secured professional autonomy and a grievance process.

What does human in the loop look like in your own practice?

In a large system, someone else picks the tool, the template and the settings. In your own practice, you hold all three. That is what the striking doctors asked for, and it is worth using.

The last row is the point. The responsibility never moves. Only the control does.

What should I ask before I trust an AI scribe?

1. Can I read and edit every line before anything is saved to the chart?

2. Can I use my own template, or only theirs?

3. Can I turn it off for one visit, or for a patient who says no?

4. Does it suggest codes, and can I turn that off?

5. If I leave, what happens to my templates and my data? Put it in the contract.

Frequently asked questions

Does human in the loop mean an AI scribe is safe?

No. It means a person can catch errors, if they have the time and the authority to fix them. Safety also depends on what the tool leaves out, where the audio goes, and who chose it.

Did the Allina doctors win protections on AI?

Not in writing. MPR News reported that the tentative agreement reached on September 24 did not include specific AI terms. It secured professional autonomy and a grievance process.

Does Solanto keep a human in the loop?

In all three senses. You choose it. Solanto will draft the history, and you will read, edit and sign every note. You will be able to switch it off for any visit, and Private Mode will keep the note on the device.

The short version

Human in the loop is three things: you chose the tool, you read every note, and you can switch it off. In your own practice, you already hold all three. Solanto will keep them with you: it will write the history in your template, leave the judgment to you, and run in Cloud mode or in Private Mode, where the note never leaves your device.

The price is on the page, not behind a call. Early access is open: solanto.ai/pricing

Send this to a colleague who is choosing a scribe this month. Ask them which of the three they are about to give up.