Who Was Eliza and Why Her Death Matters
Eliza is a historical AI program created in the mid 1960s at MIT, often cited as one of the earliest examples of natural language processing. Its creator, Joseph Weizenbaum, designed it to simulate a psychotherapist by using simple pattern matching rules. The program itself did not die in a modern corporate sense, but its active development and public relevance faded as more advanced systems emerged. Today, references to "when did Eliza die" usually point to the moment its influence waned and newer models took over. Understanding this timeline helps contextualize the evolution of conversational AI and the current dominance of large language models.
While Eliza never had a corporate death certificate, its last major public deployment in research settings occurred in the early 1970s as interest shifted to more robust expert systems. The program's source code and documentation remain accessible in academic archives, serving as a foundational reference for AI historians. Its conceptual death is tied to the transition from rule based chatbots to machine learning driven dialogue systems that began in the late 1980s and accelerated in the 2010s. This shift is well documented by institutions tracking AI milestones and technology adoption curves.
Key Dates and Corporate Context Around Eliza
Eliza was first implemented in 1966 using the MAD SLANG programming language on an IBM 7094 computer at MIT. The program gained significant media attention after Weizenbaum published a description in a 1966 MIT technical report, which is often cited in modern AI retrospectives. By the early 1970s, the program was no longer at the cutting edge of research as more sophisticated systems like SHRDLU and later expert systems captured the spotlight. The exact moment of its obsolescence is not a single date but a gradual decline through the 1970s and 1980s as hardware and algorithms advanced.
Timeline of Decline
The decline of Eliza as a relevant AI tool accelerated with the rise of personal computing in the 1980s and the internet in the 1990s. Early chatbot successors such as ALICE and Jabberwacky built on more flexible architectures that moved beyond the rigid pattern matching of the original system. In the corporate world, companies like IBM and later Google invested heavily in statistical language models, making rule based systems like Eliza obsolete for commercial applications. The final chapter of Eliza's active life is often placed in the late 1970s, when it was retired from active research use in favor of more scalable approaches.
Modern Relevance and Legacy of Eliza
Today, Eliza lives on as a historical reference point in AI education and as a symbol of early natural language understanding. Its simple architecture is often recreated in programming tutorials to teach concepts like pattern matching and state machines, and it remains a popular entry point for students learning about the history of chatbots. The question "when did Eliza die" continues to generate search interest because it connects to broader discussions about AI lifecycle, hype cycles, and the rapid obsolescence of technology. Modern large language models from companies like OpenAI and Google have rendered the original Eliza approach functionally extinct in practical applications, even as its conceptual legacy endures.
The legacy of Eliza is also evident in how current AI companies frame their products as breakthroughs, often overlooking the decades of incremental progress that preceded them. Regulatory bodies such as the SEC and organizations like the Partnership on AI regularly reference early AI systems like Eliza when discussing the governance of modern generative models. Understanding when and why Eliza faded provides a factual baseline for evaluating the longevity of today's AI startups and large language model providers. This historical context is crucial for investors, researchers, and policymakers assessing the sustainability of current AI trends and the potential for future systems to follow a similar trajectory of rapid rise and decline.