The Guardian’s AI chief questions technology companies for using news without paying: “In Silicon Valley they don’t understand what journalism is”

“In Silicon Valley they don’t understand what journalism is”. The phrase was fired by Caspar Llewellyn Smith, head of artificial intelligence at the British newspaper The Guardian, in a criticism of companies that use news to feed your products without paying for the work behind it. If a response from ChatGPT or Claude replaces a visit to the outlet that researched and published that information, how is the next news funded?

Llewellyn Smith visited the newsroom this Wednesday Clarionwhere he met with editors and managers of the newspaper, before his presentation this Thursday at Media Party. During his stay he addressed how AI is being used in journalism, but also and above all, how it is affecting it.

The motto of this edition of the Media Party, “The agent is the message” (reformulation of the well-known phrase of the communication theorist Marshall McLuhan) was the starting point of the interview. Agents are artificial intelligence systems capable of performing tasks autonomously, from searching and analyzing information to interacting with applications and executing actions on behalf of a user: in journalism they became a kind of assistant to the reporter or editor.

But a big problem is that the way those tools answer queries It also affects media income. “Google Zero” describes a scenario in which the search engine stops sending traffic to other sites: a user searches for information and receives a summary made with AI, without having to open the pages from which that information was extracted. Media lose visits, advertising revenue and opportunities to add subscribers, while your work continues to feed search engine responses.

This whole situation reached the courts of the United States, with the lawsuit by The New York Times against Microsoft and OpenAI. In parallel, other media outlets are negotiating licenses and The Guardian is one of those that made agreements, precisely with OpenAI, which once again raises a tension that the media already went through during the social media boom: how to negotiate with technology companies without giving up control over their work and access to their readers, as happened with Google and Facebook.

Before his talk at the meeting that brings together journalists, programmers, designers and entrepreneurs in Buenos Aires to work on media innovation, he spoke hand in hand with Clarion.

Google Zero: what happens when the search engine stops sending readers

—How did AI agents change journalistic work?

—There are very clear opportunities within the newsrooms. If you are an investigative journalist, you can put several agents to work to search the web or examine large data sets. They can read thousands of documents in a time that would take a person weeks, months or years. In that sense, they can give them a kind of superpowers to journalists, I am quite optimistic about that potential.

─And what worries you about the use of AI in journalism?

─What worries me is what happens if Bill Gates is right. Three years ago he said that we would all have an AI personal assistant with technology far superior to that available then. It’s easy to imagine that world: an agent who knows a lot about you, gets the information you need and anticipates your requests. There will be people who continue to enter a media outlet because they trust its brand and the human criteria that selected the news, but for basic queries, agents can already do a great job. What is today a small Google summary can become much richer and more personalized: If consumption moves there and replaces the mediathe challenge for the media is very great.

—How would you explain the “Google Zero” phenomenon and what problems does it pose for journalism?

—In the era of paper, the media found stories, published them and controlled their distribution. With the Internet they were able to reach more people, but they gave up a good part of that access to Google and Facebook. But now AI introduces another change: it takes the raw material of the news, synthesizes the journalism, interprets it, puts together a response and presents it to the public. When Google answers a query directly, many people are satisfied and do not need to go to the middle, it doesn’t click: That reduces advertising revenue and the chances of converting those readers into subscribers. It is emptying the business model.

─In fact, global traffic to media and websites in general is falling.

─Of course, and the media that is too dependent on Google is suffering. Specialized information or a strong brand can protect you. But I wouldn’t assume that that protection could last three, four or five years. AI assistants will know their users very well and offer increasingly better answers. There will be fewer reasons to visit the person who produced the news.

—If an AI can explain a topic or summarize current events, what does journalism continue to contribute to civil society?

—Two days ago I was talking to our director about re-watching investigations like the Panama Papers and the Paradise Papers, to see if an AI finds something we hadn’t detected. An AI cannot yet speak to the survivor of an attack, understand their experience, and provide the human empathy necessary to tell that story. Knocking on doors, asking difficult or sensitive questions – that’s not something you’re going to replace tomorrow. There is also human value in opinion and commentary, in trusting the judgment of a human.

Who pays for the news that chatbots use

—Journalists are poorly paid around the world and, at the same time, the media lose advertising revenue. If AI uses that work without bringing readers to the sources, how is the production of information sustained? What incentives are there to do journalism?

—The first thing I would say is that there is no scenario in which, in five years, journalists will be better paid. And I would distrust an ecosystem that incentivizes them excessively: people do journalism because they are motivated by something more fundamental, which is discovering things. That’s the sad truth. That said, it is possible to be positive, because you would think that an assistant acting on your behalf could find the best source on politics, local information or football and explain why they consider it reliable. And that source is going to be “you”: a specialized journalist. If it is aligned with our interests, it could increase demand for specialized journalism and help it find larger audiences.

─In other words, the demand for more journalism may increase.

─Well, the information revolutions increased demand: with the printing press, people wanted to read more books. With AI, a query can be much more detailed than the few words we typed into Google. The question is how those who produce that information are paid: That’s going to be the fight. If technology companies do not find a mechanism, they will also have a problem: they will run out of new data for their models.

—Of course, Big Tech needs journalism to feed their products. Cloudflare, for example, proposes a model where media outlets are paid for using their content. Do you see that model as viable?

—Let’s see, AI needs three things: algorithms, computing power and data. A star Silicon Valley developer can cost the same as many people in our newsroom. These companies spend fortunes on developers and computing. But without the third pillar, data, they cannot make their models work. They became accustomed, since the early days of the Internet, to thinking that they could take this material for free.. Some behaved and continue to behave appallingly. We saw Anthropic have to pay for stealing works from authors. And we saw Meta try to change their terms to say that if you have an Instagram account, they can take all your journalism and content and do whatever they want with it. It’s outrageous.

─How do we get these companies to have so many licenses?

─Some of them behave this way because they are used to thinking that this material is free and they do not understand journalism. In Silicon Valley they don’t understand what journalism is. Nor do they seem too interested in having people doing investigative journalism, something that is not so surprising when we talk about the richest companies in the world. How do you convince them that what we do has value?

─Well, Jeff Bezos said, when he bought the Washington Post, that journalism gave away its content for years on the internet.

─There are signs from Google, Apple and Microsoft that they will have to reach a fairer agreement. The media we must promote rules and standards that facilitate this negotiation. A better result is possible and would be naive assuming that technology companies put the interests of society first: that is not necessarily the story of the last thirty years.

—The Guardian already has an agreement with OpenAI. Isn’t it a double-edged sword like trusting Google and SEO was?

—We made licensing agreements, for example with ChatGPT. They pay us and we authorize them to use our journalism to synthesize it and answer a question. What they can’t do is give you the full article within ChatGPT. Today we still think of it as a reader referral model. If you ask what one of our columnists thinks, the system can summarize it and put a link for you to read the note on our site. The problem is that many people are not going to leave the chatbot: There is an additional step and the answer may already be enough for you.

—How could it be another model?

—The New York Times puso un paywallwe ask readers to support us voluntarily. Today most of our business comes from them. What if you could access the Guardian within ChatGPT, Gemini, or another assistant and read an entire article? From the point of view of a medium it is not the most desirable future, but it could be the least bad alternative if we preserve part of the relationship with that reader. If someone is subscribed to Clarín, why couldn’t they use that subscription within ChatGPT or Gemini? That’s not built yet. There are versions with other services, such as Spotify. I am interested in exploring whether it can be an additional model to licensing.

—AI companies often play the “Skynet card” when they warn of the dangers of their own creations. Beyond this marketing narrative, what is the specific threat to journalism?

—The biggest threat is that the power concentrates on a shrinking number of Silicon Valley companies and that they become the usual way in which we all access the digital world. If news is something that these companies synthesize, organize and return to consumers, and those who produce that knowledge do not have a way to get paid, The media industry is going to be emptied.

─What is the biggest threat to society?

─There is a risk that these systems could be corrupted. In an increasingly authoritarian world, it is no longer absurd to worry about a scenario in which your assistant knows you very well, you receive all your news through him, and yet, is telling you things that are not true or distorting the truth.

─What should we demand from these companies?

─As a society, we have to ensure that these attendees represent our interests and are not manipulated. Don’t let them start introducing covert advertising or suggesting who to vote for when the elections approach. Although there are elements to think that all this can be avoided, we can also look at a company like Facebook, which has not managed our digital lives well. These are companies that did not perform well in the past and whose incentives are not aligned with ours. That should worry us.

Media Party has been held since 2012 in Buenos Aires. The current edition runs from October 29 to 31, with its first two days at the Konex Cultural City.

By Editor

One thought on “The Guardian’s AI chief questions technology companies for using news without paying: “In Silicon Valley they don’t understand what journalism is””
  1. vesinhxanhsac的个人资料 – 襄阳好工作 – Powered by Discuz!
    vesinhxanhsac的个人资料
    提示信息 – 杰成电气自动化 – Powered by Discuz!
    vesinhxanhsac的个人资料 齐鲁热线
    vesinhxanhsac的个人资料 – 平望新博电脑,苏州博之宏智能科技有限公司 – Powered by Discuz!
    vesinhxanhsac的个人资料 小扑网
    vesinhxanhsac的个人资料 – 黑喵制造总局 – Powered by Discuz!
    vesinhxanhsac的空间 – ABCDV网站 – Powered by Discuz!
    vesinhxanhsac的个人资料 大同法制网
    Giặt nệm bao lâu khô và cách kiểm tra trước khi dùng
    Ve Sinh Nem Cho Gia Dinh Co Tre Nho Checklist Thuc Te
    Rệp giường và vệ sinh nệm: Cần xử lý ra sao?
    HedgeDoc – Collaborative markdown notes
    HedgeDoc – Collaborative markdown notes
    HedgeDoc – Ideas grow better together
    Giặt nệm tại nhà khu Nam TP.HCM: Chuẩn bị thế nào?
    Giặt nệm tại nhà TP.HCM: Cách chuẩn bị và lưu ý
    New
    Giặt nệm tại nhà khu trung tâm TP.HCM: Cách chuẩn bị
    HedgeDoc – Collaborative markdown notes
    HedgeDoc – Collaborative markdown notes
    Người bị dị ứng cần lưu ý gì khi vệ sinh nệm?
    Bao lâu nên giặt nệm một lần? Lập lịch theo tình trạng — vesinhxanhsachdep
    HedgeDoc – Collaborative markdown notes
    HedgeDoc – Collaborative markdown notes
    HedgeDoc – Collaborative markdown notes
    HedgeDoc – Collaborative markdown notes
    Vệ sinh nệm cho người bị dị ứng: Điều nên lưu ý — filtenborgmortensen48426
    HedgeDoc – Collaborative markdown notes
    HedgeDoc – Collaborative markdown notes

Leave a Reply