PDFs are difficult for AI to analyze due to their structure, which could lead to the development of alternative formats that are easier to read.

The PDF format, created by Adobe in 1993, is extremely popular for documents, but poses a challenge for artificial intelligence systems. Although PDFs are designed to be read by humans, their complicated structure, such as columns, graphics, and tables, hinders automatic analysis. AI programs must use optical character recognition (OCR) technologies to extract text, but they encounter difficulties with complex documents. In contrast, formats like HTML are easier to analyze due to the tags that indicate the structure of the document.

The limitations of PDFs affect both users, who have difficulties in analyzing them, and AI companies, which cannot access a large volume of data stored in unstructured formats. In this context, startups like Factify are exploring the development of new formats that combine the advantages of PDF with easier analysis by AI. Additionally, the company Mistral is working on an AI-based OCR system to improve the reading of PDF documents. Although PDF remains the dominant standard, the demand for more machine-friendly formats is growing as artificial intelligence becomes increasingly relevant.