In a stunningly short space of time, ChatGPT, Gemini and a host of other generative AI plat-forms have become part of our everyday lives. These AI-driven search developments stand on the shoulders of years of advancements, shown in the timeline below. The majority were out of sight, driving the now-ubiquitous global engine that is Google.
In the 2010s, deep learning advanced rapidly, notably through the rise and development of transformer models (“neural network” technology inspired by how our brains work). This transformed image recognition, language modeling and the interpretation of user queries by search engines, significantly improving how they understand and respond to the words people use in their searches. Large-scale transformer models like OpenAI’s GPT-4 have made significant progress in generating human-like text and understanding context, dramatically influencing the current world of search and setting its future direction.
Expectations from here include:
• More precise and contextually aware search results driven by enhanced AI-powered developments in advanced natural language processing (NLP), a branch of AI that helps computers to comprehend, generate and manipulate human language.
• The integration of quantum computing, revolutionizing search algorithms and dramatically boosting performance results.
• Immersive and augmented reality, and multimodal capabilities—integrating text, voice, image and video inputs—will transform how and where users interact with information online.
• The already significant private/corporate use of Large Language Models spearheads a push for ethical AI practices to prioritize privacy and data security.
1990 The First Search Engine—Archie (archive without the ‘v’) was created by Alan Emtage, a student at McGill University, and publicly released on September 10, 1990. It indexed FTP archives, making it easier to find specific files.
1993 The First Web Crawler: World Wide Web Wanderer—created by Matthew Gray, it was used to measure and track the growth of the web.
1994 Yahoo!—originally a directory of websites organized by category.
• Infoseek—known for its very complex system of search modifiers.
• WebCrawler—the first engine to index entire web pages, not just titles, created by Brian Pinkerton at the University of Washington.
• Lycos—one of the earliest and most popular web search engines known for the combination of directory- and text-searching capabilities.
1995 AltaVista—innovative search engine renowned for advanced features and comprehensive web page indexing.
• Excite—hybrid web directory and search engine known for portal features and multimedia capabilities.
1996 Backrub—the precursor to Google—developed by Larry Page and Sergey Brin at Stanford University, it used backlinks to rank the importance of web pages.
• HotBot—known for its innovative features like customizable search filters.
1997 Ask Jeeves—allowed users to ask questions in everyday language; made notable advancements in understanding natural language.
1998 Google—PageRank algorithm revolutionized web search by ranking pages based on their importance derived from inbound links; quickly became the dominant search engine.
2004 Personalized search—Google begins personalizing results based on users’ search history and preferences.
2005 Google Maps—local search capabilities integrated with maps, providing geographic context to results.
2009 Bing—Microsoft’s search engine, known for its visual search capabilities and integration with other Microsoft services.
2010 Semantic Search capabilities, Neural Networks, Machine Learning and AI—applications including image recognition and natural language processing. Advanced search algorithms from simple word-matching to sophisticated systems interpreting various types of context.
2011 Google Panda update— notable improvement to query results reducing the ranking of low-quality sites and increasing the ranking of high-quality content.
2013 Google Hummingbird update—first algorithm designed to understand the human intent behind a search query.
2015 Launch of RankBrain—a machine learning-based component of Google’s algorithm that helps process and understand search queries.
2017 Development of Transformer Models—type of neural network architecture that learns context by tracking relationships in sequential data like sentences.
2018 Google introduces BERT—a significant advancement in NLP, Bidirectional Encoder Representations from Transformers, or BERT, help Google understand the sentiment, context and paraphrasing of words in search queries.
2020 OpenAI releases GPT-3—one of the largest AI models to date, having a significant impact on the ability of search engines to generate human-like text and better understand context.
2021 Google launches Multi-task Unified Model (MUM)—designed to access all media formats to collect information and further enhance the results for complex search queries.
2022 OpenAI launches ChatGPT—an AI system able to have conversations; Microsoft integrates ChatGPT into Bing for chat-based search queries; Google launches Bard, their AI chat system for search.
2024 Google releases Circle—a program that translates highlighting, tapping or circling an image, video or text into a list of search results.