LAION (acronym for Large-scale Artificial Intelligence Open Network) is a German non-profit which makes open-sourced artificial intelligence models and datasets.[1] It is best known for releasing a number of large datasets of images and captions scraped from the web which have been used to train a number of high-profile text-to-image models, including Stable Diffusion and Imagen.[2][3]
In February 2023, LAION was named in the Getty Images lawsuit against Stable Diffusion as a non-party.[4] In April 2023, LAION was directly sued by a German photographer who wanted to have his images removed from the training set.[5] In September 2024, the Regional Court of Hamburg dismissed the lawsuit, in what was described as a "landmark ruling on TDM [Text and data mining] exceptions for AI training data" in Germany and the EU more generally.[6]
On April 15, 2023, LAION and contributors publicly released an open source AI assistant chatbot called OpenAssistant.
Image datasets
LAION has publicly released a number of large datasets of image-caption pairs which have been widely used by AI researchers. The data is derived from the Common Crawl, a dataset of scraped web pages. The developers searched the crawled html for <img> tags and treated their alt attributes as captions. They used CLIP to identify and discard images whose content did not appear to match their captions.[7] LAION does not host the content of scraped images themselves; rather, the dataset contains URLs pointing to images, which researchers must download themselves.[8]
The first such dataset, LAION-400M, was released in August 2021 and consisted of 400 million image-caption pairs. The pairs were extracted from a random subset of webpages scraped by Common Crawl between 2014 and 2021.[9] It was an attempt to recreate the process used by OpenAI to collect the 400 million image-caption pairs they used to train the CLIP model - the company had chosen to open-source the model's code and weights, but not its training dataset.[7]Imagen, a text-to-image model announced by Google Brain in 2022, was trained on LAION-400M in combination with private internal datasets.[10]
A successor of more than 5 billion pairs, LAION-5B, was released in March 2022.[11] As of its release, it was the largest freely available dataset of image-caption pairs in existence.[7] Its creation was funded by Doodlebot, Hugging Face and Stability AI, the AI company behind the funding of the Stable Diffusion text-to-image model, which was trained on it.[12]
Criticism
Several studies show that the images in LAION-5B contain problematic images and text pairs of rape, pornography, malign stereotypes, racist and ethnic slurs, and other extremely problematic content.[13][14]
An investigation by Bayerischer Rundfunk showed that LAION's datasets, hosted on Hugging Face, contain large amounts of private and sensitive data harvested from public websites.[15]
In December 2023, the Stanford Internet Observatory released a report on LAION-5B that found 3,226 suspected instances of links to child sexual abuse material with 1,008 of these being externally validated. In response, LAION temporarily removed LAION-5B and LAION-400M citing its "zero tolerance policy for illegal content" and "an abundance of caution".[16] In August 2024, LAION released a cleaned dataset called Re-LAION-5B.[17]
OpenAssistant is an artificial intelligence (AI) open source chat-based assistant that understands tasks, can interact with third-party systems and retrieve information dynamically to do so. The project is developed by a group of volunteers in collaboration with LAION. One of the goals for development includes free access to large language models that can be run locally on consumer hardware.[18][19] The project is backed by a worldwide crowdsourcing effort involving over 13,500 volunteers who have created 600k human-generated data points.[19][20]
^Birhane, Abeba; Prabhu, Vinay; Han, Sang; Boddeti, Vishnu Naresh; Luccioni, Alexandra Sasha (2023-11-06), Into the LAIONs Den: Investigating Hate in Multimodal Datasets, arXiv:2311.03449
