Perplexity’s Unethical Practices Threaten Web’s Foundations, Raising Concerns for AI Industry’s Integrity

Perplexity, an AI startup creating a Google Search competitor, has been engaging in unethical practices to build its “answer engine,” raising questions about the company’s integrity and the broader implications for the web.

Key issues with Perplexity’s approach: Perplexity’s practices have come under scrutiny for their potential to undermine the web’s foundations and the work of primary sources:

• Perplexity’s “answer engine” acts as a rent-seeking middleman, providing answers directly to users rather than sending traffic to the original sources, thus depriving them of ad revenue.
• The company’s Pages product creates entire aggregated articles by plagiarizing content from primary sources, going beyond simply quoting a sentence or two to answer a user’s query.
• Perplexity has been caught dodging paywalls and barely citing original investigations, as well as using copyrighted art without permission.

Perplexity’s response and further revelations: When confronted about its practices, Perplexity’s leadership offered weak justifications and failed to commit to ethical solutions:

• Perplexity’s CEO, Aravind Srinivas, blamed third-party scrapers for ignoring the robots.txt code that asks web crawlers not to scrape pages, but declined to name the scraper or commit to terminating the contract or convincing the scraper to honor the code.
• The company plagiarized an article about its own plagiarism, despite the publication explicitly blocking Perplexity in its robots.txt file.
• Perplexity has been found to surface AI-generated results and misinformation, calling into question Srinivas’ commitment to “factfulness.”

Unethical practices and broken trust: Perplexity’s actions highlight a growing trend of AI companies engaging in questionable practices to obtain data for their models:

• Srinivas himself bragged about lying to Twitter by pretending to be an academic researcher to scrape data for Perplexity, demonstrating a willingness to deceive for the company’s benefit.
• Perplexity’s reliance on abusing third-party policies and shattering the trust that has built the internet raises concerns about the company’s true value proposition and the integrity of its leadership.

Broader implications for the web and AI industry: Perplexity’s practices have far-reaching consequences for the future of the web and the AI industry as a whole:

• The company’s actions threaten the good-faith agreements, such as the robots.txt code, that have held up the web for decades, potentially setting a dangerous precedent for other AI companies to follow.
• As Perplexity seeks to raise hundreds of millions of dollars from investors, the question remains whether its users and backers will hold the company accountable for its unethical behavior or turn a blind eye in pursuit of profit and growth in the AI space.