Channel 1 Will Use Generative AI to Hyper-personalize Your News. An AI-Enabled CNN?
Handwringing over the future of news aside, Channel 1 faces long odds of success
Channel 1 pre-announced its AI-enabled news service that will arrive in 2024. The news will be brought to you by virtual human newscasters in multiple languages. News scripts will be written by large language models (LLM) as summaries and then edited by humans before being recorded by virtual humans for user consumption.
The editorial coverage will focus on global news, finance, and entertainment, and the company promises to enable users to access the content in any language. Users will also be able to choose their preferred virtual human anchor and its personality. According to a Channel 1 virtual human anchor:
Everything you will see on 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.
Channel 1 says it will launch a mobile app in 2024 that will learn user preferences and customize the news experience. This will reopen the debate over the relative value of personalized versus curated content. It will also create new friction with news organizations around other companies' use of their content and whether the work is transformational or illegally republished.
A Focus on Personalization
While this is not the first company to use virtual human newscasters or employ generative AI for script writing, Channel 1 may be the first to apply the technologies for mass personalization of news. Synthedia’s generative AI benefits model includes concepts such as hyper-automation and hyper-creation and also notes the opportunity for hyper-personalization. Because generative AI can create novel outputs, it has the potential to move us beyond today’s mass market information and entertainment models that cannot afford to deliver one-to-one personalization.
Granted, I am not convinced Channel 1’s personalization will be any different than you get from YouTube. Company co-founder Adam Mosam told The Hollywood Reporter that the company intends to “produce 500 to 1,000 segments daily.” With a pre-approved news inventory, Channel 1 will simply match segments with user interest. This is different than creating a bespoke newscast based on your personal financial portfolio, which may include segments that you are the only audience member who will consumer it.
Generative AI makes this level of personalized news curation possible, but Channel 1’s plan to have humans in the editorial loop will preclude it. Still, personalization works. YouTube’s chief product officer, Neal Mohan, said during a CES panel discussion in 2018 that over 70% of the time spent on the streaming video platform was due to AI-driven recommendations. Channel 1 may be able to provide a similar news and lifestyle experience based on a more highly curated set of content.
Who will create original content?
“Founder and entrepreneur Adam Mosam said the news aired on the network will come from legacy outlets and commissioned freelance reporters. Additionally, the AI will generate its own reporting from public records and government documents,” wrote USAToday’s Anthony Robledo.
The news sourcing and reviews by human editors are the key methods Channel 1 says will ensure stories are factually accurate. According to one of Channel 1’s virtual human anchors from the video:
What is AI-native news. First, here is what it is not. It’s not fake news. There isn’t a computer somewhere writing its own stories about things that haven’t happened. Our system relies only on trusted news sources to bring together the best journalism on any topic from around the world into a single news program.
The company also plans to use generative AI to create illustrative examples of news and events that were not captured visually. Channel 1 says it is like a newsroom sketch where cameras are not allowed. However, this is a false analogy. The courtroom sketch artist is in the room and depicts the events as they interpret them. Channel 1 would be taking descriptions and creating something new without any direct knowledge of how they looked visually.
While the company says these images will be labeled as AI-generated, they definitely will be criticized as misleading. I suspect this will be the most controversial idea and would not be surprised if it is ultimately dropped altogether or severely curtailed.
The Future of Media
It is fair to ask how different this is from news anchors that read information directly from a teleprompter. Some anchors write their own scripts, but often, others write all or part of the stories for them. In those instances, there may be little difference between the news anchor and the virtual human. The anchors are merely vessels for delivering information. In both cases, the news organization, writers, and editors are the accountable parties. That will be the same for Channel 1.
The bigger issue is sourcing news. If news distributors continually crowd out organizations that conduct original reporting, will there ultimately be even more scarcity in source information than we see today? Frankly, I don’t see this as any different than what we see today on the web. There are thousands of news site and blogs that do little or no original reporting. They merely republish information that originated elsewhere. In fact, even the big news organizations that conduct original reporting also re-report stories created by others.
This suggests legacy new organizations and creators of original news content are not at as much risk as they might want you to believe. Channel 1, if successful, is more likely to initially draw the audience away from non-traditional news and blogs that do not create original content.
There was a time when USA Today was “derided” and its approach questioned by the media establishment. The Internet Archive (AKA, WaybackMachine) saved a story that Editor and Publisher Magazine online some time ago removed. Referring to USA Today on its 30th anniversary, John K. Hartman wrote:
The paper was derided by critics as “McPaper” for serving up nuggets of information rather than fully informing the public, as traditional newspapers purported to do…
USA Today became the country’s most circulated and best-read newspaper with more than 5 million daily readers. It cut into the readership of nearly every regional newspaper in the country and forced them to grudgingly change their ways and be more responsive to readers, especially younger readers who were more television oriented and were beginning to reject traditional newspapers.
CNN also faced criticism when it first aired, though much of that was written before the internet archived stories. However, it is accepted that CNN changed news media coverage and consumption by creating a continuous, real-time news cycle.
Similarly, the rise of blogging led to a new round of establishment media criticism of a novel form of news publishing. J.D. Lasica wrote in the Fall 2003 edition of Niemen Reports:
Many traditional journalists are dismissive of bloggers, describing them as self-interested or unskilled amateurs.
Channel 1 is unlikely to radically challenge existing media but it may serve as a catalyst for change. Axel Springer cut a deal with OpenAI to serve up its news content to ChatGPT users while also using generative AI technology to make newsrooms more efficient.
South Korea’s MBN News introduced a digital twin of one of its lead anchors in November 2020. Chinese Central Television introduced “AI anchor Wang” in 2022 to deliver “news and economic programs,” according to The Korea Herald. DeFiance Media introduced a novel AI news anchor and a replica of a technology futurist in 2022. Change is already underway.
Channel 1’s Prospects
However, there is no guarantee that Channel 1 will find an audience. News is a crowded category and news aggregation apps have been unable to achieve breakout or lasting success. Flipboard and SmartNews both generated initial interest, and the former even peaked at 145 million monthly active users (MAU), while the latter has attracted tens of millions of mobile app downloads across iOS and Android. However, neither has seriously challenged the business models of established linear media organizations or digital new media companies.
Apple News and Google News are exceptions. However, the companies do not create their own content, and news is an ancillary activity. Apple and Google are simply leveraging their distribution advantage to drive more platform activity. Similarly, Flipboard, SmartNews, and other startups are not creating original news content. Channel 1 will transform news instead of simply linking to it. In this way, it will compete more with ChatGPT than existing applications.
It will also directly compete with the likes of CNN, Sky News, and other televised news organizations and indirectly with all other news publishers. The question is whether it heralds the arrival of the next CNN or USA Today in terms of news industry disruption. Alternatively, it may be the next Quartz, which began with a lot of hype but never found its footing in a hypercompetitive new landscape.
I think a Quartz trajectory is more likely than a CNN-like disruption by Channel 1. Its success will not be driven by avatars but through personalization. If it can crack personalization and achieve the scale of hundreds of millions of daily active users, it has a chance to build and sustain a loyal user base. The information it would gain in that scenario would help it establish a competitive moat. Without both ingredients, it will look like repackaged news easily copied by new competitors and existing media organizations, several of which are desperate to find a new operating model.
Channel 1’s biggest impact may be to give legacy news publishers a scapegoat to justify implementing the AI-enabled automation they already want to deploy in the name of productivity. And it won’t be just Channel 1. Before Channel 1 launches its mobile app in 2024, you will likely see at least two more startups claiming to have a similar model. Channel 1 has no moat, but it has a great video that should help raise venture capital funding.