ChatGPT Is Turning the Internet Into Plumbing
What does life online look like filtered through a bot?
![OpenAI's logo on top of a pipe](https://cdn.theatlantic.com/thumbor/d_eMMBIb334FTiDhkuFpBIxiA2s=/0x0:2000x1125/960x540/media/img/mt/2023/12/chat_gbt_plumbing_1/original.jpg)
There is a tension at the heart of ChatGPT that may soon snap. Does the technology expand our world or constrain it? Which is to say, do AI-powered chatbots open new doors to learning and discovery, or do they instead risk siloing off information and leaving us stuck with unreliable access to truth?
Earlier today, OpenAI, the maker of ChatGPT, announced a partnership with the media conglomerate Axel Springer that seems to get us closer to an answer. Under the arrangement, ChatGPT will gain the capacity to present its users with “summaries of selected global news content” published by the news organizations in Axel Springer’s portfolio, which includes Politico and Business Insider. The details are not altogether clear, but the announcement indicates that when you query ChatGPT, the bot will be able to spin up responses based on Axel Springer stories, accompanied by links to the stories themselves. Likewise, material from Axel Springer publications will be used as training data for OpenAI, advancing the company’s products—which may have already consumed something like the entire internet.
It’s arguably a strange move for the publisher, which in the old days might have seen some competitive advantage in maintaining a distinctive voice—that is, one that isn’t easily replicated by a chatbot. But Axel Springer will be paid for lending its work. That’s certainly better than getting ripped off for free, which is effectively what generative AI is presumed to have done to publishers across the industry. Julia Sommerfield, a spokesperson for Axel Springer, declined to give any specific details about the deal but told me over email, “Our reporters at Politico and Business Insider will continue to deliver high-quality journalism. The partnership introduces an additional channel for distribution and revenue, and enriches users’ experience on ChatGPT.” A spokesperson for OpenAI declined to comment on the deal.
It makes sense that the press release uses the phrase “global news content”—content is an ugly but useful word for understanding exactly what’s happening here. At its core, generative AI cannot distinguish original journalism from any other bit of writing; to the machine, it’s all slop pushed through the pipes and splattered out the other end. For this reason, the deal is notable not just for media wonks, but because it says something about the future of the internet—in particular, the vision that OpenAI has for it.
OpenAI’s most powerful model does not currently provide information about any event more recent than April 2023. That will change: Although OpenAI has an agreement to use archival material from the Associated Press, Axel Springer is reportedly the first publisher to provide ongoing news stories in this way. The benefit of this arrangement, in theory, is that current, accurate info is instantly available, and can conform to exactly what a ChatGPT user wants to know. But the generative-AI era has introduced a distancing effect. ChatGPT, Bing, or Google’s Gemini may present readers with information and links from publications, but they hardly seem to incentivize engagement with those publications. If ChatGPT reproduces news updates, what reason is there to click on the original? How many times have you Googled for more information after reading a headline on an elevator or taxicab screen?
The shift to news via chatbot feels ironic when you consider what generative AI has wrought elsewhere: Old-school search engines such as Google have been flooded with supercharged, optimized spam and are struggling to handle it, while sites such as CNET and Gizmodo have published deeply flawed synthetic writing in a desperate bid to stay competitive. ChatGPT is becoming more capable at the same time that its underlying technology is destroying much of the web as we’ve known it.
Until such time as bad AI content can be quickly identified and dealt with, its sheer volume will continue to crowd out legitimate sources. This will be fine to the extent that people can still find good information; major publishers are likely to endure, and surely more of them will sign deals with OpenAI. But it is a sour development for the overall diversity of the web. For a long time, the internet was about discovery, about jumping from site to site to find different perspectives and styles; it was, in some sense, an equalizer or a democratizer. That became less true as we began to experience the internet through social-media platforms that serve as gatekeepers; it is becoming less true still as new generative-AI infrastructure is built on these ruins.
That seems to be the direction all of this is headed in. All of the websites, all of the writing: It’s plumbing for a digital faucet. With fresh content surging, ChatGPT will become more viable as a one-stop shop through which to experience the internet. It will offer something that better resembles the full range of human knowledge as it exists at any given moment—albeit dotted by “hallucinations,” lending just a bit of doubt to every interaction.
However flawed it is now, perhaps this transformation will ultimately be understood as an expansion of human potential. There are persuasive arguments that generative AI inspires creativity and facilitates good work, which is more than could be said of most websites. Whatever the case, it is certainly the end of one era and the beginning of another—one that will not be defined by the panoply of our digital creations but instead by a chatbot’s text box and its cursor, blinking and blinking, awaiting your command.