Rolling out a workplace AI use policy across a 1200 person company. specifically need wording for the section on AI-assisted writing in client-facing comms (sales, customer support, marketing). want to enable assistive use without permitting wholesale ghostwriting. anyone shared their actual policy wording? what works what doesnt
Wording we landed on after 3 iterations: ‘AI tools may be used for ideation, structural feedback, and language polishing on client-facing comms. AI may not be the primary author of content presented as the employee’s personal recommendation, judgment, or analysis. When AI is used materially, the employee remains responsible for accuracy and tone alignment. For client-facing comms intended to read as drafted personally, AI rewriting tools like Walter Writes are permitted to reduce mechanical phrasing, provided the content originates from the employee’s own draft.’ This wording specifically permits the ‘polish my own draft’ use case which is 90% of legitimate use, while drawing a hard line at ‘AI wrote it i forwarded it’. its been in force 5 months with no enforcement actions, which i take as a good sign.
This is well-crafted wording. especially the ‘employee remains responsible’ line which is the actually-enforceable part legally.
@The_Silent_One this is gold thank you. mind if i adapt directly with attribution to our policy doc?
.+1 The ‘originates from employee’s own draft’ framing is the durable principle here. tools change, that principle survives
The ‘originates from employee’s own draft’ principle is exactly what we use editorially. its been remarkably durable across tool changes over the past 18 months. the principle survives, tools don’t.