Perplexity AI bị tố

Perplexity AI, a well-funded startup backed by Jeff Bezos and Nvidia, has become one of the most controversial companies in the AI space. Its search engine promises synthesized, cited answers to user queries, but the way it obtains and reproduces those answers has triggered a cascade of legal and ethical challenges that have continued into 2026.

This article examines the core ethical issues facing Perplexity: copyright infringement and plagiarism, unethical web scraping, academic integrity risks, and the problem of AI-generated misinformation and hallucination. It also covers the company's attempt to address these concerns through a revenue-sharing program and why that program has not stopped the lawsuits.

Copyright Infringement and Plagiarism: The Dominant Issue

The most significant ethical charge against Perplexity is that it systematically copies and reproduces copyrighted content from news publishers without proper permission or attribution. Multiple major media organizations have publicly accused the company of plagiarism .

Forbes discovered that Perplexity's "Perplexity Pages" feature published a plagiarized version of a subscriber-only article. The copied content included directly replicated phrasing and an illustration taken from a previous Forbes article, yet the original source was barely credited with a small 'F' icon at the bottom of paragraphs . Forbes' Chief Content Officer Randall Lane described the practice as "cynical theft" .

Wired reported that Perplexity likely scraped its site and other Condé Nast properties "thousands of times" despite those sites having used blocking measures. The publication found that Perplexity's AI summaries reproduced large sections of their articles nearly verbatim . The investigation even found that when Wired published a story about how Perplexity is a "bullshit machine," Perplexity then plagiarized that same story .

A Growing Wave of Lawsuits

The plagiarism accusations have translated into a steady stream of legal actions:

• Condé Nast (parent company of Wired, The New Yorker, and others) sent a cease-and-desist letter accusing Perplexity of plagiarism and bypassing the Robots Exclusion Protocol used by websites to block scraping .
• The New York Times filed a lawsuit in December 2025, claiming Perplexity unlawfully replicated millions of articles and that its AI produces hallucinated content incorrectly attributed to the Times .
• Reddit sued Perplexity in October 2025 for harvesting user posts to train its AI model, accusing the startup of scraping copyrighted material by masquerading its bots as ordinary users .
• News Corp (owner of The Wall Street Journal, The Times, and others) also sued Perplexity for "massive scale" copying and repurposing its content without authorization .
• CNN reportedly joined the legal fray in 2026, raising questions about copyright and the future of journalism .

Investigations by AI content detector Copyleaks found that Perplexity was able to paraphrase or plagiarize substantial portions of paywalled articles, even content it claimed not to have access to .

Unethical Web Scraping

A related ethical issue is Perplexity's approach to web scraping. The startup has been accused of ignoring the Robots Exclusion Protocol, the long-standing standard that websites use to deny or grant permission to automated crawlers . Investigations revealed that Perplexity's web crawlers continued to access restricted content from publishers who had explicitly opted out .

This behavior has been characterized as a violation of the understood norms of the web, even if specific legal rulings on its legality remain in flux .

Academic Integrity Risks

Beyond the world of news publishing, Perplexity has raised concerns in academic settings. A study on the use of Perplexity AI in academic writing found that students risk indirect plagiarism when they copy AI-generated results without editing or understanding the content, violating principles of academic honesty . Researchers note that while AI can help compile writing, the responsibility for the authenticity of the work remains with the user, and the tool may generate information or quotes from unclear sources .

Misinformation and Hallucination

Investigations have also found that Perplexity's AI answers sometimes include fabricated content alongside plagiarized material. Wired reported that when asked about unique stories the AI shouldn't have had access to, the chatbot returned summaries that were inaccurate and merged real information with made-up details . The New York Times lawsuit similarly alleges that Perplexity's outputs contain AI "hallucinations" that are then wrongly attributed to the newspaper .

This combination of copying real content and inventing false details presents a unique danger: users may trust the answers because parts of them are verifiably real, making the fabricated portions harder to detect .

The Revenue-Sharing Response

In response to the wave of accusations, Perplexity introduced a revenue-sharing program for publishers in July 2024 . The program was announced as a way to share advertising revenue with content creators. However, the program has not resolved the underlying disputes. Cease-and-desist letters and lawsuits—from Condé Nast, The New York Times, Reddit, News Corp, and others—have continued well into 2026 .

Legal experts cited in multiple reports note that the fundamental tension is that the closer a Perplexity answer gets to a publisher's original expression, or the more it replaces the need for a user to visit a publisher's site, the more it invites copyright and unfair-competition claims . The question is no longer theoretical: it has become a central product-defining issue for AI search companies.

The Bottom Line

The ethical problems facing Perplexity AI are not minor or isolated incidents. They represent a systemic conflict between the business model of AI-powered "answer engines" and the rights and economic incentives of content creators. The company faces multiple high-stakes lawsuits, public accusations from some of the most prominent publishers in the world, and growing scrutiny of its web-scraping practices. Whether Perplexity's citation-based approach can survive legal challenges, or whether it will ultimately be reined in by courts or regulation, will have significant implications for the future of AI search.