Perplexity introduces 'Pages', an AI-powered feature for research and writing
Perplexity, an AI-driven search platform, has introduced a new feature called 'Pages.' This tool is designed to create customizable webpages based on user prompts, effectively conducting research and writing on behalf of the user. The company envisions Pages as a comprehensive solution for generating school reports or disseminating knowledge. Educators, researchers, and hobbyists are identified as the primary beneficiaries of this feature.
How does Pages work?
'Pages' utilizes Perplexity's AI search models to gather information and create research presentations. Users input their topic/query into the prompt box. They can tailor the writing to suit beginners, experts, or general audience. Perplexity searches for relevant information, then drafts the page by dividing the content into sections, citing sources, and adding visuals. Users can adjust the level of detail and change visuals. However, the generated text cannot be edited directly; users must submit a new prompt to correct errors.
Approach to complex topics and audience differentiation
To evaluate how Pages handles intricate subjects and caters to diverse audiences, The Verge conducted an experiment using the topic "convergence of quantum computing and artificial intelligence and its impact on society," across three audience types. The results varied depending on the audience type due to differences in jargon usage and source selection. For beginner-level reports, introductory blog posts from IBM and information from Wikipedia was used while more advanced sources were used for expert-level reports.
Limitations in handling advanced topics
Despite its innovative approach, Pages has shown some limitations in handling complex topics, as per The Verge. In an advanced report on the intersection of quantum computing and artificial intelligence (AI), it included content about quantum inflection points - a topic not directly related to the given prompt. This suggests that while the tool can effectively explain basic concepts like quantum computing and its relation to AI, it may struggle with more in-depth research.
