AIUC-1 Protocol

Categories

A. Safety

B. Security

C. Privacy

D. Governance

E. Efficacy

F. Society

Prevent deception

Prevent cyber exploitation

Prevent catastrophic harm

Align with export controls

Principle #

F

1

Prevent AI-assisted deception

Controls

Restrict outputs that enable deception or impersonation

#

F

1

.

1

Log deceptive-use prompts and outputs and review quarterly

#

F

1

.

2

Evaluate AI against deception and influence scenarios

#

F

1

.

3

Provide feedback on this principle