A startup with aspirations of becoming a trusted global news brand is planning to launch a service delivering curated content – written by AI and delivered by artificial anchors.

Channel 1 demoed its tech with a half-hour [1]preview of its AI-generated news on Twitter this week, with AI anchors describing the startup's plans and reading some sample news items.

(It's causing some of us to remember [2]Ananova from the year 2000, an over-hyped computer-animated-and-voiced news anchor run by a UK newswire. She was later [3]flogged off to a telco and ultimately dumped.)

[4]

The LA-based virtual station claims all news it presents will be fact-checked by humans to ensure accuracy before being placed in the virtual mouths of its artificial newsreaders.

[5]

[6]

The content on [7]Channel 1 "relies on trusted sources and fact checking and uses AI to give you news the way you want it – personalized, localized and distilled," one of the virtual talking heads recited, with another trying to dispel fears by noting that "journalism's core values of integrity and accountability are at the heart of everything we do."

Channel 1 plans to source news from a variety of places – including several big-name news agencies that founders Adam Mosam and Scott Zabielski told The Register in an interview today they'll make public next month. The photo-channel also plans to source stories from a network of vetted independent journalists, and eventually wants its AI to be able to formulate stories from fact-based documents like government reports, SEC filings and court cases. The pair told us AI is particularly good at generating news from simple fact-based items like those, but it will still use human editors to clean the stories up.

[8]

Along with adapted and remixed stories, the channel will also use AI to generate images to "create footage of events where cameras were not able to capture the action." One of the AI anchors likened the use of AI images to courtroom sketches that "aren't a literal depiction of actual events but can still provide important information and nuance to a news report." Channel 1 promised all AI-generated imagery will be labeled as such.

As for the news itself, sources will be cited on Channel 1 in a manner similar to how they're cited on existing news channels, we're told – with text in the corner indicating whence it came.

"We're starting with news vetted from trusted sources," Zabielski stated, adding that the startup wants to earn audience trust by delivering trustworthy news free of AI hallucinations. "We're only going to get one chance to do this right," Zabielski added.

[9]

The ultimate goal, Mosam and Zabielski explained, is to launch an app that delivers the aforementioned personalized news feed. But Channel 1's first iteration will be a free ad-supported streaming television (FAST) channel – essentially the digital version of a live TV station, where a few hours of news repeat on a loop.

[10]Don't be fooled: Google faked its Gemini AI voice demo

[11]Sports Illustrated boss fired – but it's totally nothing to do with AI fake news

[12]Publisher breaks news by using bots to write inaccurate stories

[13]China cracks down on AI-generated news anchors

The pair hope to launch the FAST channel in spring 2024, all delivered by soulless, dead-eyed avatars – though to be fair the "AInchors" (yes, really) are slightly more believable than that.

"The first phase of our launch was passing the uncanny valley and delivering something that's a believable thing to watch," Mosam told The Register – which the preview reel demonstrated. While the quality of the various AInchors that appear on the screen varies to a degree – with some using hand gestures that don't quite line up with their speech, and a few moments of stiff movement and out-of-sync audio – the star anchor in the video, a middle-age blonde woman, was fairly believable.

[14]

Nope, not a human – Click to enlarge

The startup, which currently has ten employees and is entirely self-funded, will need to scale to succeed, both founders admitted. The amount of content it can produce will be constrained by both its technology and the amount of fact-checking, re-writing and editing that can be done by each person in a single day. Channel 1 is in talks with investors right now and plans to begin larger fundraising efforts soon.

And we had to ask the question: Do we really need fake anchors – albeit ones that Channel 1 wants to develop with a sort of digital "personality" that can allow them to report from their own vantage point – reading regurgitated news that's tailored to each user?

Someone has to do it, Mosam told us. "This is coming … and we're doing our best to stand in front of a speeding train in the most responsible way possible," he declared.

"There will be plenty of cases where there's improper use, fraud, fake news and bad actors," Mosam predicted. "What we're trying to say is that we can be a responsible choice." ®

Get our [15]Tech Resources



[1] https://twitter.com/channel1_ai/status/1734591810033373231

[2] https://www.theregister.com/2000/04/19/ananova_speaks_cyberrubbish/

[3] https://www.theregister.com/2002/05/20/ananova_pushes_mobile_scraps_email/

[4] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_offbeat/bootnotes&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=2&c=2ZXwx1BFYancHB1hCMqrSWgAAAIQ&t=ct%3Dns%26unitnum%3D2%26raptor%3Dcondor%26pos%3Dtop%26test%3D0

[5] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_offbeat/bootnotes&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44ZXwx1BFYancHB1hCMqrSWgAAAIQ&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0

[6] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_offbeat/bootnotes&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33ZXwx1BFYancHB1hCMqrSWgAAAIQ&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0

[7] https://www.channel1.ai/

[8] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_offbeat/bootnotes&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44ZXwx1BFYancHB1hCMqrSWgAAAIQ&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0

[9] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_offbeat/bootnotes&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33ZXwx1BFYancHB1hCMqrSWgAAAIQ&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0

[10] https://www.theregister.com/2023/12/11/ai_in_brief/

[11] https://www.theregister.com/2023/12/12/sports_illustrated_ceo_fired/

[12] https://www.theregister.com/2023/01/19/cnet_reviewing_ai_authored_stories/

[13] https://www.theregister.com/2023/05/16/china_crackdown_on_ai_generated_news/

[14] https://regmedia.co.uk/2023/12/14/channel-1-news-sample.jpg

[15] https://whitepapers.theregister.com/