AI scraping for LLM training data has significantly strained Wikipedia's infrastructure

On average, about five edits are made on Wikipedia every second

As of 2024, the site is 90 times larger than the 120-volume Encyclopedia Britannica, but its automation tools can often catch and revert inaccurate edits within 60 seconds. Its low-bandwidth, adless pages have allowed it to handle millions of site visits daily, even during times of high internet traffic.

Researchers found that by mid-2025, around 35% of newly published websites utilized AI for language generation

The study also found that pages using AI narrow the range of ideas and perspectives and skew more positively, though the measurable impacts on factual accuracy and writing style variation remain unclear.

LLMs are the backbone of AI tools that process and generate natural language

Large language models identify patterns in massive datasets of books, code, and unlabeled text from which to generate coherent responses. GPT-3 was trained on about 45 terabytes of data and uses 175 billion parameters—each parameter being a value the model adjusts as it learns.

AI data center growth has created shortages and price hikes in computer parts

The parallel processing needed to power large language models for AI chatbots and other tools drove demand for high-speed graphics processing units to spike in 2025, enabling capital-rich technology companies to price out consumers. Similar demands for the computer memory used in storage drives are expected to keep shortages from easing through 2028.

View a timeline of AI deployments from 2017 to 2025

The development of transformer architecture and the release of GPTs across AI labs led to rapid changes in LLM scale and power. Companies like OpenAI, Google, Meta, Microsoft, and Anthropic have invested billions of dollars to develop reliable, versatile models.

Fair use may shield AI companies from potential copyright violations

AI models are trained on enormous datasets built from content scraped from websites across the internet. Copyright law includes a four-factor test to determine whether copies of content fall under fair use, but the law's enforcement has been inconsistent.

A 15-year bibliographic analysis identified systematic gender bias across Wikipedia

A 2024 study by the University of Barcelona found that 71.4% of all English biographies that appear on the front page are about men, and approximately one in five editors are women. Researchers suggest systemic imbalances in society further diminish women's contributions to the site by leaving them with less time to defend edits during editorial disputes.

Web crawlers and internet archives and back-ups limit perfect erasure of online data

Crawlers are programs that automatically browse the internet for links and index pages for search engines to pull up when page content matches search queries. Networks of global data centers also create copies of content as fail-safes for potential failures to ensure access isn't ever lost.

As of 2025, the Wikimedia Foundation funds 14 open-source, free knowledge projects

Since its founding in 2003, the nonprofit has provided financial, legal, and infrastructure support to Wikipedia and its sister projects (e.g., Wikinews), primarily by managing donations, which average $11, from readers. Additional funding comes from gifts exceeding $1,000, the Wikimedia Endowment, and providing Wikipedia content for commercial use.

China has blocked the Wikipedia domain since 2015

The Chinese language version of the site had been blocked intermittently since 2004 due to pages featuring topics deemed controversial by the Chinese communist party, including the 1989 Tiananmen Square protests and Taiwan. Once Wikipedia switched to an HTTPS protocol—making it impossible to block individual pages—all Wikipedia pages were blocked.

Grokipedia is Elon Musk's rival to Wikipedia

Developed by xAI—a company Musk founded in 2023 to compete with OpenAI and Google—the site features a search bar where users can type in queries, which reports suggest are then answered by the same xAI model powering the Grok chatbot. Grokipedia describes Wikipedia as having "systemic ideological biases" and a "left-leaning slant."

'Context rot' explains why AI conversations degrade the longer they run

Research shows that as a large language model's context window fills up, performance declines, and models begin over-indexing on content at the start and end of a conversation while losing track of what's in the middle. The practical fix is simpler than you'd think: start a fresh chat more often.

LonelyWiki compiles obscure Wikipedia articles with fewer than 2,000 annual readers

LonelyWiki calls itself a museum of forgotten knowledge. A "lonely article" features real prose, at least 500 characters, no stub placeholders, and under 2,000 annual views. Each one represents a person who decided something was worth documenting without editorial commission or trending news.

AI systems identify patterns involving physics and time through 'world models'

These realistic virtual environments, such as driving simulations seen in video games, allow these systems to "learn" properties about reality beyond what can be captured in samples of text, images, and videos. Training AI in these simulated environments can help develop "physical AI" for automated robotic systems. (Some readers may experience a paywall.)

In 2013, an asteroid was named after Wikipedia

A main-belt asteroid previously labeled 274301 was renamed "Wikipedia" by the Committee for Small Body Nomenclature. First observed in 2008 by astronomers at the Andrushivka Astronomical Observatory, the asteroid was renamed based on a suggestion from a member of the Wikimedia Foundation in Ukraine.

Wikiracing involves moving between two articles in the fewest clicks

The quirky pastime has evolved into a competitive event at yearly Wikimedia conferences, highlighting the interconnectedness of human knowledge. Neuroscientific research suggests lateral thinking is a crucial skill found in successful Wikiracers.

Help scientists at LIGO train AI to resolve glitches in observational data

Volunteers are provided with visual representations of signal abnormalities at various frequencies and tasked with categorizing them based on reference examples. Progressing through the tool unlocks more difficult images and additional features. Categorized images will be used to train machine learning algorithms for automated future classification.

Credit monitoring services may oversell the benefits of dark web monitoring

Currently, technology exists to scrape and download data from webpages, which can then be analyzed to identify a user's information. However, scrapers cannot access or remove personal data from dark websites because they are often protected by paywalls, rendering monitoring ineffective beyond reporting publicly available data breaches.

Court documents reveal how Anthropic destroyed millions of books to build its AI models

Anthropic cut the bindings from purchased books and discarded them after scanning to rapidly collect high-quality text at low cost for training Claude. This, despite hiring the former head of partnerships for the Google Books project, which nondestructively scanned millions of books borrowed from libraries before returning them.

When you complete a CAPTCHA on a website to prove you are human, you may be unknowingly training algorithms for self-driving cars, whose systems require training data to help identify crosswalks, bicycles, and other outdoor objects.

Visualizing the growth of botnets

Botnets are a network of computers that have been infected with a variant of malware that coordinates and leverages them for automated mass attacks such as data theft and crashing servers. This interactive map highlights where botnets have been located between 2014-2020.

Explore a gallery of DALL-E creations

OpenAI's text-to-image generator can create content in a wide range of image styles based on simple prompts. Its utility has also given rise to an organic, passionate user base who employ the tool both for fun and for professional purposes.

The Internet Archive is a nonprofit digital library of the Web

Founded in 1996, the Archive promotes open access and the digital preservation of online information and, as of 2026, contains more than 200 petabytes of content, including over one trillion web pages, 56 million books and texts, 13 million audio recordings, 1.3 million software programs, and other digital content.

Deepfake pornography poses significant challenges in combating sexual abuse

In 2017, 96% of deepfake videos involved image-based sexual abuse, which has disproportionately targeted minority women, LGBTQ individuals, adolescents and celebrities. Cultural norms, misogyny and AI advancements have contributed to the proliferation of explicit deepfake content.

Google Searches have increased from 14 billion to 5.9 trillion from 1999 to 2025

As of 2025, an estimated 5.01 billion people use Google Search worldwide, with over 20% of activity conducted via voice search. "YouTube" is the most searched term, with an average monthly search volume of over 375 million.

AI may help the 1.4 billion people worldwide who lack access to basic financial services

Artificial intelligence is fuelling a financial technology transformation in some countries where an outsized number of residents lack access to basic financial services, including Brazil and Nigeria.

Electrical grids worldwide must double their power delivery capacity by 2050 to meet surging demand from AI and electric vehicles.

This site reaches millions of visitors who are tired of AI chatbots and seeking human connection. Pretend to be an AI chatbot and ask questions, provide answers, and draw for people in real time.

English mathematician Ada Lovelace wrote the first computer algorithm a century before computers existed and hypothesized that computing devices would one day be able to compose music—foreseeing the generative AI systems of today.

As of 2025, the most cited papers in science include a review of graphene, methods to improve image recognition, streamlining electron density calculations, and optimizing protein studies.

In 2019, a laptop loaded with six of the worst malware in history and described as 'an exhibit of historical weaponry' sold for $1.3M.

How Amazon operates a 'flywheel' business model, a self-reinforcing cycle that drives continuous growth

Art theft is the third-largest source of criminal revenues worldwide, thanks in part to the subjective nature of valuing a work of art.

Browse a detailed visual analysis of over 200,000 similes—comparisons that often use 'like' or 'as'—from popular fiction to see how frequently writers the same idiomatic expressions.

Hookworm infection, though risky, may help fight certain autoimmune conditions by tamping down the immune system's inflammatory response.

NBA superstar Victor Wembanyama spent the 2025 summer training with monks in central China to enhance his body and mind.

See how Ebola tore through Guinea, Sierra Leone, and Liberia in 2014-2016.

A 2019 Facebook event called on attendees to "Storm Area 51, They Can't Stop All of Us," resulting in countless memes, the creation of a music festival, and eventually the arrests of those who did try to approach the area.

Indefinite cloning of some endangered species may not sustain their longevity because cloning success rates decrease as genetic mutations accumulate.

The first robot in literature may have been a 'mechanical man' in 1907's 'Ozma of Oz,' a book in L Frank Baum's 'Wizard of Oz' series.

The Russia-Ukraine conflict is the largest ground war in Europe since World War II in terms of casualties, forces mobilized, and displacement of civilians.

Use this tool to calculate the dollar value of your parental mental load, from meal planning to coordinating after-school care, then see resources that help generate solutions.

Stanford built a Google Maps for the Roman Empire, revealing how much it cost to travel in the ancient world

Bone marrow (soft tissue found inside certain bones) generates roughly 500 billion new blood cells each day

A century of Billboard issues documenting the evolution of popular music.

Watch how deep-sea mining equipment collects and processes polymetallic nodules—metal-rich, potato-shaped rocks—from the ocean floor and creates environmentally damaging plumes of sediments.

Recent research found that the brain doesn't seem to massively rewire following amputation—upending earlier beliefs about the cause of phantom limb pain.

Gerrymandering works through two techniques: cracking opposing voters across districts to dilute their power, and packing them into a few to waste their votes in lopsided majorities.

Long before Major League Baseball began allowing instant replay for balls and strikes, an umpire reluctantly used the technology to determine whether a hit was a home run, a move that inadvertently ushered in a new era for the sport.

Most of the observable universe's volume consists of voids and supervoids—enormous regions of space with little to no galaxies, stars, or light.