ELIZA, a chatbot developed in the 1960s and recognised as the world’s first, has been revived using long-lost computer code found in archival records. Originally created by MIT professor Joseph Weizenbaum, ELIZA was designed as a conversational program capable of mimicking a psychotherapist’s interactions. The project, conducted by researchers and archivists, marks a milestone in understanding early artificial intelligence and its impact on modern technology. Despite its simplicity compared to today’s AI, ELIZA’s ability to hold a conversation remains impressive.

Reconstruction of the Code

According to a paper posted on the preprint server arXiv, the chatbot’s code was uncovered in 2021 by Jeff Shrager, a cognitive scientist at Stanford University, and Myles Crowley, an MIT archivist. Written in a now-obsolete programming language called MAD-SLIP, the original 420-line code had not been operational for six decades. The research team spent years debugging and creating a computer emulator capable of running the software. ELIZA’s functionality was successfully restored on December 21, 2024.

Preserving Historical Authenticity

As reported by Live Science, researchers encountered a bug in the code but decided against fixing it, citing the need to preserve its historical integrity. Shrager explained to Live Science that altering the program would compromise its authenticity, likening it to modifying an iconic artwork. This decision highlights the importance placed on maintaining the original program’s features, even at the cost of usability.

Impact and Legacy

Experts emphasised ELIZA’s significance in shaping the development of artificial intelligence. David Berry, a digital humanities professor at the University of Sussex, stated that while modern language models surpass ELIZA in capability, its conversational design remains noteworthy. The chatbot was programmed to listen and prompt users, a feature seen as more conversationally authentic than many current AI systems.

The revival of ELIZA draws attention to the need for preserving the history of computer science, as its legacy is considered a cultural artifact reflecting the early days of computational innovation.