{
  "path": "/human-generated-content-2",
  "site": "at://did:plc:xgvzy7ni6ig6ievcbls5jaxe/site.standard.publication/3mgfwckliwc2m",
  "tags": [
    "human generated content",
    "fediverse",
    "generative-ai",
    "activitypub",
    "chatgpt",
    "social media",
    "openai",
    "llm"
  ],
  "$type": "site.standard.document",
  "title": "Human-Generated Content 2",
  "description": "Publishers are seeing two very different futures for their businesses. Is the future of media aggregated and summarized or is it direct-to-audience?",
  "publishedAt": "2024-06-06T18:37:46.000Z",
  "textContent": "Hello, again! \n\nThe first edition of this series brought the largest amount of new subscribers augment has ever seen, so I want to start by thanking all of you for joining me on this journey. As you can probably tell, I'm still trying to figure this out but I hope I'm leaving you with a few nuggets of knowledge and context with each post I leave in your inbox. \n\nBefore we jump in: I also wrote two other pieces on augment in between issues. Check them out if you haven't yet, I'm really happy with how they turned out:\n\n- Google Danced - Will Apple?\n- Patreon Belongs in the Fediverse\n\nFor this edition, I want to talk about two conversations happening around the future of journalism. One side sees the end-times and is handing their content to be aggregated and summarized for easy consumption. The other side also sees the end-times, but is opting to change the way they distribute and build their audience. Yup - it's another AI and Fediverse conversation. But I also warned you it would be last time so I hope it isn't a surprise.\n\nSo let's look at these two forks in the road and see what their potential for the future is.\n\nThe Publisher's Dilemma \n\nI want to start this with a blog post by Ben Werdmuller (Senior Director of Technology at ProPublica) that summarizes the current social media news landscape, different business models that have been attempted in the past, and the various solutions ahead of them.\n\n> We’re at the foothills of both changes: social media is likely to fragment further, and generative AI will become even more prevalent as it becomes more powerful. Newsrooms can no longer rely on their old tactics to reach their audiences, and they will need to build new tactics that take these trends into account if they hope to survive.\n\nSo here we are: the distribution strategies of the 2010s are looking a whole lot like the newsstands of the mid-2000s. The first time, the newsstands scaled back because the audience changed the way they consume the news. This time, however, it's because the newsstands themselves have either decided they don't want to sell the news or that they would rather summarize the day's news instead - sometimes lying to you confidently as they do.\n\nThere's a lot of more in Ben's piece and I urge you to read it, along with subscribing to his newsletter. His experience in the industry has taught him about the dynamics between journalism and technology and I've learned a lot from reading his work. \n\nI want to add another description from the piece:\n\n> These organizations must own their digital presence outright to shield themselves from risks created by third parties that wield outsized supplier power over their business models. They must build direct relationships with their communities, prioritizing open protocols over proprietary systems. They need to invest in technology expertise that can help them weather these changes and make that expertise a first-class part of their senior leadership teams.\n\nSo, which one are publishers choosing - open protocols or proprietary systems? Well, it's complicated.\n\nPublishers Turn to AI Companies\n\nLet's start with the publishers opting into proprietary systems. Namely, in the last few weeks Vox, The Atlantic, and News Corp have all signed deals with OpenAI. These deals go two ways - OpenAI gets a pipeline of up-to-date news content to add to their models and publishers get deals on OpenAI tools, some of which will actually be really useful for both the publisher and its readers.\n\nIt is important, however, to note that how news content will surface on ChatGPT is still unclear once these deals have a tangible impact on the experience. When I ask for the day's news on the platform today, this is what it looks like:\nOpenAI responds with what it considers the four top news stories of the day\nFirst off, these are definitely not the biggest news stories of the day—there's an election happening in India and a Bollywood actress's endorsement is certainly not the most important part of it. But I digress.\n\nAnother thing to note about this response is the attribution: it does link me to news sites where some of this is sourced but it doesn't take me to the article, it takes me to the homepages. There are no quotes from the pieces and no authors called out. There's four stories with three sites and it doesn't tell you which stories line up with which publications. Now, maybe all this will change after these deals but I currently see no real benefit to the media businesses other than an upfront, short-term bank deposit. \n\nBut it somehow gets worse. The next day I got more specific about what news I want to focus on and ChatGPT just straight up lied to me:\n!Me: Tell me about yesterday's elections in India. ChatGPT: The 2024 Indian general elections concluded with the counting of votes on June 4. These elections were conducted over seven phases from April 19 to June 1. The final results declared that the Bharatiya Janata Party (BJP) secured a significant victory, solidifying its position with a majority in the Lok Sabha.  The BJP-led National Democratic Alliance (NDA) won a decisive majority, with the BJP alone securing 303 seats. This result marked another strong performance by the BJP, led by Prime Minister Narendra Modi, who is set to continue his tenure​ (IndiaTimes)​​ (Free Press Journal)​.  In addition to the national elections, state assembly elections were also held in Andhra Pradesh, Odisha, Arunachal Pradesh, and Sikkim. The BJP made notable gains in Odisha, winning 78 seats in the 147-member state assembly, marking a significant political shift in the state​ (India Today)​.  Overall, the election saw a high voter turnout and reinforced the BJP's dominance in Indian politics, both at the national and state levels.  Me: This isn't true, the BJP did not win a majority ChatGPT: The 2024 Indian general elections resulted in a hung parliament, with no single party securing a clear majority. The Bharatiya Janata Party (BJP), led by Prime Minister Narendra Modi, did not achieve the expected majority. Instead, the opposition coalition known as the INDIA bloc made significant gains, creating a more balanced power distribution in the Lok Sabha.  The BJP and its allies won around 240 seats, falling short of the 272 needed for a majority. In contrast, the INDIA bloc, a coalition of various opposition parties, secured approximately 220 seats. This outcome has led to a highly competitive and uncertain political landscape, with both blocs needing to negotiate with smaller parties to form a government​ (Hindustan Times)​​ (India Today)​.  In the state assembly elections, the BJP saw mixed results. They managed to secure a win in the Odisha state assembly with 78 seats, marking a significant political shift in the state​ (India Today)​. However, the overall national outcome has created a scenario where coalition politics will play a crucial role in the formation of the next government.ChatGPT thinks that the BJP won the majority in the 2024 Indian elections. They did not.\nThis time it does a marginally better job of attributing the news to its sources, but fails at actually being accurate about the most important news coming out of the Indian elections. Am I supposed to trust ChatGPT as a source of up-to-date truthful journalism? Am I supposed to click on the links, go to the homepage of the news site, then search for the sourced story to confirm what ChatGPT says? Why are media brands trusting this system in its current state? I have so many questions.\n\nIf we ignore the truth - which is a really weird thing to ignore - there's another major problem here. What I see here is the further consolidation - or as Kyle Chakya would say, flattening - of the media industry. ChatGPT didn't attempt to look at multiple different news sites for the same story to get multiple perspectives, they chose at most two. In other words: unlike the ten links in Google Search or even the \"Full Coverage\" option in Google News which enables the user to get multiple viewpoints, ChatGPT is choosing a couple sources based on who made deals with them.\n\nNot to get political here, but if a news story on ChatGPT is going to be sourced from a News Corp entity, I really want it to look for different perspectives. If that's where OpenAI's information is coming from, I likely won't be trusting it. I think Vanessa Otero - Founder/CEO of Ad Fontes Media, the organization that invented the Media Bias Chat - said it best:\n\n> The rise of misinformation was mostly from bad actors and extreme partisans polluting news and political information. It increased the work required to know what was true in the news, which made a lot of people not trust any news.\n\n> AI summaries are mostly from enormous tech companies, polluting every kind of information, even the banal, which increases the work required to know what is true anywhere, which is making a lot of people not trust any information. This is worse.\n\nIt also looks like the journalists on the ground floor aren't all that happy about this decision as well. Vox reporter Bryan Walsh had swift pushback for his employer:\n\n> But while I’m not a media executive, I have been working in this business for more than 20 years. In that time, I’ve seen our industry pin our hopes on search engine optimization; on the pivot to video (and back again); on Facebook and social media traffic. I can remember Apple coming to my offices at Time magazine in 2010, promising us that the iPad would save the magazine business. (It did not.) \n\n> Each time, we are promised a fruitful collaboration with tech platforms that can benefit both sides. And each time, it ultimately doesn’t work out because the interests of those tech platforms do not align, and have never fully aligned, with those of the media. But sure — maybe this time Lucy won’t pull the football away.  \n\nTell us how you really feel, Bryan.\n\nThe Verge's Editor-in-Chief, Nilay Patel had a bit of a different - or indifferent - reaction to",
  "canonicalUrl": "https://augment.ink/human-generated-content-2"
}