Issue #40 · Monday, September 28, 2026 · 6 stories

🎙️ Tokyo Rules On A Cloned Anime Voice Wednesday 🎬 Google Vids Makes 1080p Video Free

Faking a voice got cheaper. Plus the gap in YouTube's voice detector and free 1080p Google Vids.

On Wednesday a Tokyo court decides whether TikTok had to take down an account narrated by what voice actor Kenjiro Tsuda says is an AI copy of his voice. It's believed to be Japan's first case over vocal identity, and TikTok's defence is that it's just a generic male voice. That defence is the one to watch, because if it wins, any clone that is close but not exact is fair game. YouTube's new voice detection has a blind spot of its own, and it sits alongside free Omni video in Google Vids, Qwen's TTS price cut, Premiere on Android and a first-and-last-frame workflow.

NEWS

Tokyo rules Wednesday on whether TikTok had to pull an AI clone of an anime star's voice · 1 min read

Kenjiro Tsuda, Nanami's voice in Jujutsu Kaisen, sued TikTok over an anonymous account whose narration he says is an AI copy of his voice. It had over 200,000 followers and made about ¥500,000 a month. TikTok calls it a generic male voice. Japan has no voice law, only non-binding guidance, so the 30 September ruling is the first test of whether a platform must act on a cloned voice.

Google Vids now generates 1080p Omni video free for anyone with a Google account · 2 min read

Gemini Omni 1.1 Flash, previously behind Google's paid plans, now runs free in Google Vids at vids.new on desktop, in 1080p. Google does not say what the free limit is, only that its AI plans buy more access. Every clip carries an invisible SynthID watermark, so treat the footage as detectable AI if a client or platform asks. Flash-Lite voiceovers in 100+ languages are coming next.

TECHNICAL

Camera moves alone didn't connect the shots. Planning first and last frames did · 3 min read

AI for Real Life breaks down the opening of the series The Life of the Lazy Mon: six key stills with locked character references, GPT-6 Astra in Codex planning the sequence, and MiniMax H3 Max generating the transitions to an existing song. The lesson: camera movement alone didn't carry a transition; designing both ends of each shot did. The video is sponsored by Artlist, which ran the generations.

ANALYSIS

YouTube's deepfake detector will listen for your voice. It won't say if it catches clones with no face · 4 min read

YouTube says likeness detection, open to Partner Program creators since October 2025, will match speaking voices alongside faces later this year, with enrolment and alerts coming to the mobile app. Enrolment still needs ID and a selfie video. The gap: YouTube hasn't said whether it flags a video that clones your voice without showing your face, the usual way voice clones get used, narration over stock or AI footage.

TOOLS

Qwen cuts TTS about 70% and adds a model that does voice, SFX and ambience in one pass · 1 min read

Alibaba's Qwen-Audio-3.1 cuts TTS about 70%, realtime voice about 85% and speech recognition up to 95%. The new TTS-Next generates voice, sound effects and background audio in one pass from text, timestamps and up to three 30-second reference clips, capped at 120 seconds per request (240 for podcasts), Chinese and English only. Compare Gemini's TTS, which doubles its price on 1 January.

Adobe Premiere mobile lands on Android free, with 4K export and no watermark, but no keyframes yet · 3 min read

Adobe's free phone editor reached Android a year after iPhone: multi-track timeline, 4K export, Enhance Audio, no ads, no watermark and no forced login, on Android 13 or later with at least 5GB of RAM. Firefly features such as Image-to-Video and Generate Sound Effects run on credits, with paid upgrades for more. Keyframes, beat detection and handoff to desktop Premiere are still iPhone-only.