‘ELIZA,’ the world’s 1st chatbot, was just resurrected from 60-year-old computer code

Scientists have revived "ELIZA," the world’s first chatbot, from long-lost computer code, and it still functions remarkably well. Using dusty printouts from MIT archives, a team of "software archaeologists" dug up the defunct code, which had been missing for six decades, and successfully brought it back to life.

Developed in the 1960s by MIT professor Joseph Weizenbaum, ELIZA was named after Eliza Doolittle, the protagonist of the play Pygmalion, who learned to speak like an aristocratic British woman. As a groundbreaking language model, ELIZA allowed users to interact with it, laying the foundation for modern artificial intelligence (AI). Its most famous script, "DOCTOR," was designed to respond in the style of a psychotherapist, often prompting users to elaborate on their thoughts. For instance, if someone typed, “Men are all alike,” ELIZA would reply, “In what way?”

Weizenbaum originally wrote ELIZA in a now-obsolete programming language he invented called Michigan Algorithm Decoder Symmetric List Processor (MAD-SLIP). However, it was quickly adapted into Lisp, a more widely used language. As the early internet grew, the Lisp version of ELIZA became wildly popular, while the original MAD-SLIP code faded into obscurity.

The 420-line original code was presumed lost until 2021, when cognitive scientist Jeffrey Shrager of Stanford University and MIT archivist Myles Crowley discovered it tucked away in Weizenbaum’s papers. Shrager, who has a keen interest in the thought processes of early AI pioneers, emphasized the importance of preserving such artifacts. “Having computer scientists’ code is as close to having a record of their thoughts,” he said.

Bringing ELIZA back to life was no small feat. The team had to clean and debug the code and create an emulator to mimic the type of computer that would have run the program in the 1960s. On Dec. 21, they successfully revived ELIZA for the first time in 60 years.

Interestingly, the team chose not to fix a bug they discovered in the code, likening it to preserving a flaw in a historic artifact. The program crashes if the user inputs a number, such as “You are 999 today,” but correcting it would have compromised the authenticity of the original creation.

While ELIZA was intended as a research platform for human-computer communication, its novelty as a chatbot quickly overshadowed its academic purpose. Even today, it is often compared to modern large-language models (LLMs) like ChatGPT. Though it pales in capability next to its contemporary counterparts, ELIZA’s ability to hold a conversation is impressive for a program written in 1965.

What sets ELIZA apart from modern chatbots, Shrager noted, is its focus on listening. Unlike today’s LLMs, which aim to complete sentences, ELIZA was programmed to prompt users to expand on their thoughts—a feature that Shrager argues makes it a truer form of “chatting.”

David Berry, a digital humanities professor at the University of Sussex and co-author of the study, stressed the importance of preserving computing history as cultural history. “We need to work harder as a society to keep these traces of the nascent age of computation alive,” he said. “If we don’t, we will have lost the digital equivalents of the Mona Lisa, Michelangelo’s David, or the Acropolis.”