Daily Scoop for PR: AI Backlash at Larian, TikTok's U.S. Spinout, and Two Brand-Safety Tests
Larian's AI backlash: message, medium, timing
Larian Studios took heat after CEO Swen Vincke discussed using generative AI during early exploration for the next Divinity project. Fans read it as a shift toward AI replacing creative roles or showing up in the shipped game.
Vincke clarified that Larian won't ship AI features or cut teams to replace them, noting the studio employs large writer and artist teams and continues to hire. The follow-up on X got sharper in tone, including: "Holy f*** guys we're not 'pushing hard' for or replacing concept artists with AI… We use it to explore things… The artists do that."
Why the message struggled
- Vagueness: "We're exploring AI" without strict boundaries invites assumptions. Fans filled in worst-case scenarios before the clarifications landed.
- Channel and tone: Reassurance posts mixed with profanity read as defensive, not steady. That undercut credibility outside the core fanbase.
- Timing: The first interview set the narrative; the follow-up chased it. Once trust dips, every word weighs more.
What to do next time (and now)
- Publish a single source of truth: a short policy that states exactly what AI will and won't touch, what remains fully human-made, and how data is handled.
- Lead with guardrails: "No AI-generated assets shipped," "No layoffs tied to AI," "Human review for any internal use." Repeat this in every channel.
- Choose the right messenger: Pair the CEO with the Head of Art/Writing for a joint note and a 2-minute video. Calm tone, zero snark.
- Back it with proof: Team counts, open roles, and a brief look at the creative process. Put the most reassuring facts in the first 30 seconds.
- Open a feedback loop: Office hours on Discord, a pinned FAQ, and monthly updates. Show you're listening, not litigating.
If your studio is testing AI internally, copy this template
- We do not ship AI-generated art, writing, voice, or music.
- No jobs will be cut due to AI; we continue to hire for creative roles.
- Any internal AI use is sandboxed, human-reviewed, and never trained on artists' proprietary work without consent.
- We will publish and maintain an AI use policy and update it quarterly.
TikTok's U.S. spinout: planning signal for 2026
TikTok agreed to spin off its U.S. business into a new entity backed by Oracle, Silver Lake, and MGX with U.S. oversight of data, moderation, and the algorithm. ByteDance keeps a minority stake, and the deal is expected to close in early 2026, easing near-term ban concerns (Axios).
For brands and creators, this suggests continuity with caveats. Expect subtle shifts in moderation and recommendations as the new entity asserts control.
- Hold course on 2025 content plans, but diversify distribution so one policy change doesn't tank reach.
- Start updating data and disclosure language now (ownership, storage, access, escalation paths).
- Prep an FAQ for legal/compliance questions you know are coming from clients and partners.
Livestream misstep: Puka Nacua's apology and the standard
Rams WR Puka Nacua apologized after mimicking a gesture recognized as an antisemitic stereotype during a livestream. He said he didn't understand its meaning and issued a direct apology; the NFL and Rams condemned discrimination and antisemitism.
- Intent vs. impact: Acknowledge harm first. Save context for later.
- Specificity matters: Name the issue, the affected community, and the standard going forward.
- Show the fix: Commit to education, clarify consequences, and outline how streams will be moderated to prevent repeats.
- Unify voices: Team, league, player, and platform statements should ladder to the same message and values.
UnitedHealth's audit play: transparency with dates and math
UnitedHealth shared results of an independent audit and 23 action plans, aiming to complete roughly 65% by end of 2025 and the rest by March 2026 amid ongoing DOJ scrutiny of Medicare billing practices (CNBC).
- Publish the register: What each action is, who owns it, how progress is measured, next milestone, due date.
- Report cadence: Quarterly updates with percent complete, blockers, and changes. No victory laps until outcomes improve.
- Proof beats prose: Tie actions to stakeholder outcomes (provider wait times, claim accuracy, patient experience).
- Third-party oversight: Keep the independent reviewer in the loop and visible.
Your next moves this week
- Audit all AI mentions across interviews, decks, and help centers. Patch vague phrases with hard lines and examples.
- Create a two-tier apology framework: immediate statement within 2 hours, fuller plan within 24 hours with corrective steps.
- Rehearse one livestream risk drill with talent and moderators. Pre-approve a halt script and an escalation tree.
- Refresh your platform dependency map. If TikTok shifts, what's plan B for reach, creators, and paid?
- Add a public "action tracker" template to your crisis kits for any audit, recall, or safety fix.
If your comms team needs a fast primer on AI concepts to write solid guardrails and explain them clearly, scan these practical resources: AI courses by job.
Your membership also unlocks:
