The controversy surrounding Perplexity AI, an AI startup accused of plagiarizing news articles without proper attribution, has once again highlighted the ongoing tensions between publishers and companies utilizing artificial intelligence. Perplexity had launched a feature allowing users to input prompts and receive well-researched AI summaries of articles from various online sources. However, Forbes flagged similarities between one such summary and its exclusive paywalled article. Since then, investigations have alleged Perplexity may have freely copied content from other prominent publishers as well without consent.
While Perplexity CEO claims a third party was at fault, the incident shines light on the broader issue of AI technologies allegedly sidestepping standards to access paywalled articles. Publishers argue their content should not be reproduced without permission, but AI proponents say limited use should be allowed to develop conversational agents. However, journalists have worked hard to research and write pieces, so outright copying raises valid copyright concerns. At the same time, defining acceptable use remains a challenge as AI continues progressing.
Publishers have utilized techniques like robots.txt files with instructions for bots and rate limiting the number of accesses to restricted pages. However, some agents ignore such protocols, and poisoning techniques intended to damage datasets raise ethical issues. As the capabilities of AI evolve rapidly, it is important all stakeholders engage constructively to find balanced solutions that respect both innovation and creative work. With open dialogue and commitment to facts from all sides, viable policies can be formulated.
