English

Usenet’s Role in the AI Revolution: A Catalyst for Advancement

AI is the newest online trend and has the capacity to revolutionize how we operate. Companies and individuals have been making use of services like ChatGPT and Dal-e to create articles, web pages, images, and even fake movie trailers by simply writing an effective prompt. But, of course, technology like this doesn’t appear from nowhere. In fact, there are decades of research that have gone into its creation. Coding, training, and implementation require a lot of work and collaboration, much of which happened on Usenet. In many ways, AI as we know it today wouldn’t exist without Usenet and the community of dedicated developers that refined the technology on various newsgroups.

Usenet as a Technology Incubator​

Usenet has always been ahead of the curve, technologically. It was created in 1979 and became one of the first Internet services available to users outside of government contractors. In fact, “Usenet” is a portmanteau of “User Network” because it was designed for regular people to be able to connect to and enjoy it. Because it was created in labs at Duke University and the University of North Carolina, the spirit of academic curiosity is built into its very DNA. The first users were technical researchers, so much of their early conversations focused on advancements in hardware and software, new techniques, and ways to streamline data processing and transfer. A lot of our modern technical landscape was developed and improved on Usenet. To this day, Usenet still attracts tech-savvy people looking to collaborate with others who share their interests, and high-level discussions happen every day on a variety of newsgroups.

AI and Usenet Have a Long History​

The earliest discussions of AI on Usenet trace back as far as 1983, only a few years after the protocol went live. The first AI-focused newsgroup was net.ai, which was changed to comp.ai when the Big 8 Hierarchies were introduced during the Great Renaming of 1987. A lot of early AI research happened or was shared directly on Usenet, particularly during the 1980s and 1990s. Many of the questions about the ethics of AI that we are debating today were originally asked on comp.ai.philosophy, including how dangerous it could be and whether using publicly available content for training should be considered theft. Usenet was also where natural language processing, the basis for modern AI systems like OpenAI, was originally devised and where the earliest development of Lisp, a programming language designed for AI, happened.

Serious Academic Research

Of course, Usenet is not just a place for amateurs to talk about their interests. Plenty of experts and professionals currently working in their fields frequent newsgroups to find like-minded users and share new ideas. And in some cases, Usenet has been the birthplace of major research outlets like the Journal of Artificial Intelligence Research (JAIR). JAIR was first announced to the world on Usenet in news.announce.newsgroups and was followed by the creation of two newsgroups. These newsgroups drew researchers from across the globe, encouraging them to publish relevant AI research in their open-access journal. In fact, JAIR was one of the first scientific journals to publish online, a fact that they remain proud of. Today, JAIR is among the largest and most well-respected journals focused on Artificial Intelligence research, publishing cutting edge papers and spurring detailed conversations both in their own newsgroups and in related ones. JAIR is still open-access, so anyone can read its findings without paywalls, and the journal is registered as a nonprofit organization. In many ways, JAIR represents the purest manifestation of the spirit of Usenet as a place where ideas develop into realities.

Bottom Line​

Without Usenet, AI wouldn’t be as advanced as it is today, nor would it be commercially viable. Everything that AI has been used for recently was started years ago, and probably spent some time in debate on Usenet. As the technology continues to improve, you can bet that people like the publishers of JAIR and the members of comp.ai will be at the forefront, developing new ways to create even more accurate models capable of learning more quickly. If you are at all interested in AI, Usenet is the place to be!