Quora’s Poe Platform Challenges Copyright Laws by Offering Paywalled Articles for Free

To go back to this article, navigate to My Profile and then look for the saved stories section.

Authored by Tim Marchman

Poe, Quora's AI-driven chatbot service, supported by Andreessen Horowitz with an investment of $75 million, offers its users the ability to access and download articles from subscription-based news sources in the form of HTML files.

When you input the link of a WIRED article discussing how the AI-driven search tool Perplexity copied one of their articles into the Assistant bot, it produces a comprehensive summary of 235 words along with a 1-MB HTML snapshot of the full piece. This file is accessible for download straight from Poe’s chatbot servers.

WIRED successfully accessed content from subscription-based websites such as The New York Times, Bloomberg Businessweek, The Atlantic, Forbes, Defector, and 404 Media by inputting the URLs into the Assistant bot's platform, making them available for download. This incident highlights the ongoing trend of the AI sector's lax attitude towards intellectual property rights, posing a significant threat to traditional revenue streams in industries such as journalism and music.

James Grimmelmann, a digital and information law professor at Cornell University, emphasized the gravity of the copyright problem in an email. He stated, "By creating a duplicate on their server, it clearly constitutes a copyright violation." (Quora, however, challenges this interpretation, equating Poe to a cloud storage solution.)

Upon being requested to provide a summary of a test website overseen by my colleague Dhruv Mehrotra, the bot failed to deliver a summary; however, it did produce an HTML file. The server logs of the website indicated that right after the Assistant bot received instructions to condense the site's content, a server labeling itself as "Quora Bot" accessed the website. This server made no effort to check the site's robots.txt file, implying that both Poe and Quora might disregard the Robots Exclusion Protocol, a commonly recognized but not enforceable web norm.

A high-ranking official in the media, whose identity has been protected by WIRED to openly talk about a legally delicate issue currently under probe by his organization, mentioned that their website noticed traffic from servers claiming to be Quora bots. This activity was spotted right after they input prompts related to particular articles into Poe’s chatbot, which, according to him, led to the retrieval of most or the entirety of the text from those articles.

"Quora's representative, Autumn Besselman, stated through an email that Poe serves as a platform where users can engage in conversations and pose questions to an array of AI-driven bots provided by external parties. Quora itself doesn't develop or train any AI models. Additionally, Poe includes a functionality that allows users to present web page contents to a bot, though the bot's access is limited to what the website itself delivers. Besselman also expressed openness to collaborating with your technical staff to ensure that any content behind a paywall remains inaccessible to users via Poe."

In his reply to an email seeking additional information, Besselman explained that the attachments feature in Poe is designed according to user instructions and functions in a way that is comparable to services for cloud storage, saving content for later reading, and web clipping tools, all of which he argues comply with copyright legislation. Andreessen Horowitz, on the other hand, did not provide a comment when asked.

Authored by Dhruv Mehrotra

Since the original text you've provided

Authored by Simon

Crafted by Angela Watercutter

Adam D'Angelo, who helped start Quora and previously served as the Chief Technology Officer for Facebook, now holds a position on the OpenAI board. (D'Angelo participated in the decision to dismiss OpenAI's CEO Sam Altman the previous year and was the singular member to be a part of a newly established board upon Altman's reinstatement.) In an interview conducted in March, almost three months following the announcement by Andreessen Horowitz that it spearheaded an investment round in Quora, D'Angelo engaged in a conversation with David George, a general partner at the investment firm, where the topic of the connection between Quora and Poe was explored.

"D'Angelo expressed a desire for maximum integration, mentioning, "Consider how Facebook and Facebook Messenger, despite being two separate entities, still share a considerable amount. In a similar vein, we envision Poe and Quora developing a closely intertwined relationship. Our goal is to infuse Poe with more of Quora's human elements. Additionally, we aim to incorporate Quora's entire dataset into the Poe bots. Moreover, we're in the process—and have already initiated some steps—towards enabling Poe's AI to produce responses that can be found on Quora."

Visitors to Poe's website are invited to engage in conversation with an array of chatbots. This selection includes bots that provide connections to a variety of extensive language models, as well as dedicated bots such as Mrstherapist, which offers support for issues like relationships, trauma, stress, and more. The Assistant bot serves as the standard choice for users.

The chatbot known as "by Poe" on the platform, which is run using technology from AI company Anthropic, is not owned by Anthropic. Instead, Anthropic clarifies that its clients are allowed to use its API to incorporate its technology into their own services in whatever manner they choose. The bot, seemingly built on Anthropic's Claude model, does not connect to the internet. Rather, it processes text that is likely supplied to it by Quora's web scraping tools, which gather information online based on user queries. When the chatbot is given a link to an article, it can generate an HTML file for download. This file, when opened, displays the article content in a web browser, offering a user experience similar to viewing a PDF document.

Media organizations have issued threats and pursued lawsuits against artificial intelligence firms, accusing them of copyright violations. For instance, The New York Times has initiated legal proceedings against OpenAI and Microsoft for alleged copyright breaches, while Forbes is said to have dispatched a cease-and-desist letter to Perplexity, charging it with deliberate copyright infringement. (OpenAI, Microsoft, and Perplexity have all refuted these accusations.) Content creators whose works were accessed by WIRED expressed outrage upon learning of these developments.

Charlie Stadtlander, a spokesperson for The New York Times, emphasized in an email that according to legal requirements and the publication's service terms, any form of copying or republishing The Times' material without first obtaining written consent is strictly forbidden.

"Those infuriating chatbots are driving me absolutely nuts," Defector co-owner and editor-in-chief Tom Ley expressed in a text message. "Defector does not approve of its valuable blog content being snatched away by a silly chatbot supported by the brainy folks at Andreessen Horowitz."

Suggested for You …

Delivered to your email: Will Knight delves into the latest developments in artificial intelligence with Fast Forward.

Within the largest undercover operation ever conducted by the FBI

The WIRED Artificial Intelligence Elections Initiative: Monitoring over 60 worldwide electoral events

Ecuador finds itself completely at the mercy of drought conditions

Be confident: Here are the top mattresses available for online purchase

Dhruv Mehrotra

Byline:

By Dell Cameron

Journalist Kim Z

Matthew Burgess

Dmitri Alperovitch

Joseph Cox,

Additional Content from WIRED

Evaluations and Manuals

© 2024 Condé Nast. All rights reserved. WIRED might receive a share of revenue from the sale of goods bought via our website, which is part of our Affiliate Partnerships with retail partners. The content on this website is protected by copyright and cannot be copied, distributed, transmitted, stored, or used in any other way without explicit written consent from Condé Nast. Advertising Choices

Choose a global website