Feb to August. Six months to get from "there's a problem" to an actual answer on what was safe to ship. That's roughly what it costs when a platform's own enterprise customer asks "is this clean" and there's no fast way to answer. Building that answer in-house is a year of engineering. Adding it is a day.

At Venture Café Tokyo, Kenki Tamiya and Kaoru walked a room of Japanese companies through the real risks of using generative AI commercially, and what happens when AI-generated content ends up in a campaign, a product or a film without anyone checking where it came from. The fix is pretty simple, and it’s called CopySight. The best part came after the talk. Our conversation with Kaoru kept going well past the session, and we’re already exploring how to bring CopySight to more teams together. A few Japanese companies in the room also want to hear more, and those conversations are just starting. Thank you to Venture Café Tokyo for hosting, and to everyone who stayed for the questions. Tokyo, more soon 🇯🇵 #CopySight #GenerativeAI #AIGovernance

What a week. CopySight spent LA Tech Week in the middle of it all, and on Thursday night we hosted an evening of our own as part of it! Tea, music, a view of the city, and some of our favorite people in AI and entertainment, with nowhere else to be. Thank you to everyone who came and logged off with us. To our team, for making it happen. To our investors and clients, for believing in what we’re building. And to the AI and tech community in LA, for showing up the way you did. Here’s to more nights like this one! 🌆

Someone's AI short got checked. After it was already out. A film's got way more surface area than one image. Hundreds of frames, hundreds of chances something gets flagged, solo project or studio budget doesn't matter. Should verification happen before release or after? Right now, it’s mostly after.

Logging off from Tech Week for a night. Thursday, somewhere in LA. DMs are open.

You can't untrain the model. Opting out won't fix what's already out there. So the check has to happen on the output, before it ships, against one standard. That's CopyScore. Mucker x Manatt panel, #LATechWeek @Techweek_

AI platforms: your enterprise customers want to know what's in the content before they sign. Tommy talked that through today with @Sony, @Taisu_Ventures, @soneium and DayOneDream at the Sony Ventures x Taisu Ventures Innovation Summit in Singapore.

Agents generate faster than any team can review. Enterprise buyers know it, and they ask about IP risk before they sign. The platforms that close those deals will have the check built into the pipeline. Back from Singapore with @Taisu_Ventures .

No need to tell more. That's CopySight. CopyScore gives any AI image or video a single score from 1 to 100: how close it sits to protected IP, with a report that shows exactly what was flagged. 💬 Comment COPYSCORE and we'll DM you free access.

What's wrong with this video? Four things, actually. The client posts it. You get the email. Run a free CopyScore at copysight.ai before handoff. Faster than round three of notes.

Responsible filmmaking with AI, according to @tommyart at Global Stage Hollywood: licensing that keeps crediting where the IP came from, and checking your AI footage for other people's characters and faces before it ships. Fair goes both ways.

8 seconds to generate. CopyScore tells you if it's clear just as fast. Clear or flagged, in seconds. Not weeks of someone reading it by eye. Try it for yourself. Comment COPYSCORE and we’ll give you free access. Know before it ships. copysight.ai

"Responsible AI" is the studio phrase of the year. At E&O renewal the insurer asks something narrower: what's in the file, and who checked it? We're on the Global Stage Hollywood panel at the TCL Chinese Theatre tomorrow to talk about exactly that. Oct 3, LA. Get your tickets: eventbrite.com/e/responsible…