US artificial intelligence (AI) startup Perplexity AI reportedly aims to launch ads on its search platform in the fourth quarter of 2024.
Sources familiar with the matter said the company backed by Amazon.com Inc. chief executive Jeff Bezos and chip giant Nvidia Corp. intends to run ads relevant to sectors such as technology, health, entertainment, food, and finance.
The San Francisco-based firm also looks to allow advertisers to sponsor relevant inquiries below answers and acquire display ads to the right of the platform or an answer provided by its chatbot.
Using cost per mille (CPM) as its pricing model for the ads, its prices are expected to be over $50, according to the sources.
Research from Massachusetts-based search engine marketer Semrush Holdings Inc. showed that the average CPM of display ads on desktops is $2.50, while mobile videos have around $11.10 CPM.
Perplexity announced in July its plans to share ad revenue with news publishers when its AI-powered search mentions their content to answer a question. Some of the strategy’s initial partners include TIME, Der Spiegel, and Fortune.
Perplexity Plans Ads Amid Plagiarism Accusations
The reported new ad program came amid allegations from prominent media outlets Forbes and Wired that the AI startup plagiarized their written works.
Forbes accused the company in June of stealing one of its articles in the beta version of its Pages feature, while Wired said the firm purportedly scraped content from its website and other news sites without permission.
Such illicit action suggested that its chatbot might have disregarded the Robots Exclusion Protocol, a web standard that keeps bots or crawlers from accessing certain areas of a site.
Offering the capabilities of a search engine and a large language model (LLM), Perplexity’s chatbot is known for providing more detailed and real-time responses.
It also differs from the LLMs of OpenAI Inc. and Anthropic PBC, which have their own models, as it uses the former’s GPT-3 model to generate answers.
The company has since denied the accusations, saying it has not scraped information as per the publishers’ requests and that it has abided by fair use of copyright laws.
It has also implemented changes to the way its Pages reference sources and updated its chatbot to improve direct source citation in generated copies.
