← sugarbushagent.com

Ethical Mission Statement

Protect the Taste exists so that maple syrup is what its label says, and so that monitoring it strengthens the people, the forest and the communities that produce it. It is the first use case of an identity and control plane for AI agents, built by Cequs Inc.

A stack of pancakes topped with blueberries, with maple syrup pooling on a white plate.

Who this is for

1. Producer control

The farm decides what an agent may do and for how long. An agent works within authority the farm has granted for a set period, and the farm can stop it at any time.

2. Farm data

Data from an operation belongs to that farm. It is shared only with the parties the farm chooses, for purposes the farm approves.

3. Honest measurement

An agent reports what it measures and how certain it is, and claims no accuracy it has not shown. An independent laboratory can check its accuracy, as is done today at wholesale locations. A person confirms a final grade.

4. Truthful origin

Buyers and consumers can check where a lot of syrup came from, backed by a farm the state already knows. The project produces evidence; regulators and industry bodies decide what to do with it. The design uses open standards (X.500/LDAP, mutual TLS) and is available on GitHub to invited reviewers and auditors.

5. Forest steward functionality

Monitoring has the potential to provide forest data, as a stretch goal. Like a hospital drawing blood for a lab analysis, a digital agent might take samples and analyze them. What to collect would follow scientific forest management from maple research and forestry labs. Today the sensor is attached to the syrup lines, so forest steward functionality needs research and development with input from experts.

6. Fairness

The tool is affordable for small operations and has no lock-in. It monitors syrup and equipment, not individual workers.

7. Indigenous nations

Maple sugaring originates with the Indigenous peoples of northeastern North America. Indigenous producers and knowledge holders are consulted in how this is built and used.

8. Security

Agents get the least authority they need, short-lived credentials, and approval by a physical security key. The enforcement point checks the agent's status and grant window on every request, so a person can revoke an agent without waiting for a credential to expire. Security defects are tracked, and defects and security test results are disclosed to qualified evaluators.

9. Accountability

A named person answers for each agent. The project says plainly what is built and what is not yet. This is a narrow agent under human supervision. It does not address the risks of advanced AI in general.

Tell us what to change

This statement is a draft for producers and other parties to react to. The survey asks about each commitment.

Prototype site: nothing here is a real production deployment.