ELIZA is a natural language processing computer program created in 1964 to 1966[1] at MIT by Joseph Weizenbaum.[2][3] Eliza was created to explore communication between humans and machines, simulating conversation with pattern matching and substitution methodology that gives humans the illusion of understanding, with no representation that can be considered actually understanding what is being said by humans and software programs.[4][5][6] The ELIZA program is originally written [7] in MAD-SLIP, pattern matching directives that contain most of its language capability provided in separate “scripts”, represented in a lisp-like representation. The most famous script, DOCTOR, simulates a psychotherapist of the Rogerian school (where the therapist often reflects back the patient’s words to the patient),[8][9][10] and uses rules, defined in the script, to respond with non-directional questions to human inputs. ELIZA is one of the first chatterbots (“chatbot” modernly) and one of the first programs capable of trying the Turing test.[11]

ELIZA’s creator, Weizenbaum, created the program as a method to explore communication between humans and machines. He was surprised and shocked, that people attribute human-like feelings to the computer program, including Weizenbaum’s secretary.[3] Many academics believed the program can positively influence the lives of many people, particularly those with psychological issues, and that it can aid doctors working on such patients’ treatment.[3][12] ELIZA is capable of engaging in discourse but it does not converse with true understanding.[13] Many early users were convinced of ELIZA’s intelligence and understanding, even while Weizenbaum’s insisted the contrary.[6] The MAD-SLIP source-code was found in MIT archives and published on various platforms, such as archive.org.[14] The source-code is of high historical interest because it demonstrates not only the specificity of programming languages and techniques at that time, but also the beginning of software layering and abstraction as a means of achieving sophisticated software programming.

MoniGarr, the author of AIGeneration.blog was a toddler in the mid 1970s who was first learning how to read. The only books of interest in their home were C programming books that were being used with a Tandy TRS-80 from the local Radio Shack in Massena NY. MoniGarr was fascinated by the adults’ conversations (a group of communication specialist friends that had served in the Vietnam War) about the ELIZA Bots they were programming using the software engineering documents & books they were using and sharing amongst themselves. Curiosity led MoniGarr to programming her first ELIZA Bots in the mid 1970s to share never ending stories that changed based on each choice the human made while text chatting with the custom ELIZA Bot scripts that MoniGarr created. In those days when MoniGarr was first learning how to read as a toddler, their ELIZA program was many if / else statements that output the creative stories MoniGarr wrote with the C programming language on the family’s Tandy TRS-80. ELIZA went on to inspire MoniGarr to create the world’s first Kanien’keha Chatbots in the mid 1990s using techniques from ELIZA and it’s descendant A.L.I.C.E. to earn a Bronze Medal for ‘Most Knowledgeable Bot’ in a world wide tech competition. MoniGarr debated that measuring / judging a chatbot’s intelligence based on it’s ability to mimic or behave as English speaking men do was flawed. The debates encouraged the competition judges to allow MoniGarr to compete with them with their Kanien’keha (Mohawk Language) chatbots.


A conversation between a human and ELIZA’s DOCTOR

Joseph Weizenbaum‘s ELIZA DOCTOR script, was created to provide a parody of “the responses of a non-directional psychotherapist in an initial psychiatric interview”[15] and to “demonstrate the communication between man and machine was superficial”.[16] While ELIZA is best known for acting like a psychotherapist, the speech patterns are from the data and instructions supplied by the DOCTOR script.[17] ELIZA examines the text for keywords, applied values to the keywords, and transforms the input into an output; the ELIZA script determines the keywords, sets the values of keywords, and sets the rules of transformation of the output.[18] Weizenbaum chose to make the DOCTOR script in the context of psychotherapy to “sidestep the problem of giving the program a data base of real-world knowledge”,[3] as in a Rogerian therapeutic situation, the program only has to reflect back the patient’s statements.[3] The algorithms of DOCTOR allows for simulated intelligent responses that some humans view as huma traits, intelligence and sentience.[19]

A working-class character in George Bernard Shaw‘s Pygmalion , named Eliza Doolittle, inspired Weizenbaum to name his software program ELIZA. Weizenbaum said ELIZA’s ability to be “incrementally improved” by various users made it similar to Eliza Doolittle,[18] since Eliza Doolittle was taught to speak with an upper-class accent in Shaw’s play.[8][20] Unlike Shaw’s play, ELIZA is not capable of learning new patterns of speech or new words through interaction alone. Edits need to be made directly to ELIZA’s active script in order to change the way the software works.

Weizenbaum first implemented ELIZA in his own SLIP list-processing language, where, depending on the first entries from the human, the illusion of human intelligence would appear, or be dispelled with several interchanges.[2] ELIZA’s responses were so convincing that Weizenbaum and several others have anecdotes of users getting emotionally attached to the program and forgetting they were conversing with a computer program.[3] Weizenbaum’s own secretary asked Weizenbaum to leave the room so she and ELIZA could have a real conversation. Weizenbaum was surprised and wrote: “I had not realized … that extremely short exposures to a relatively simple computer program could induce powerful delusional thinking in quite normal people.”[21]

Interactive computing (via a teletype) was new in 1966. It was 15 years before personal computers became familiar to the general public and three decades before most people encountered attempts at natural language processing with Ask.com, Internet services or PC help systems such as Microsoft Office Clippit. Those programs include years of research and work, while ELIZA remains a milestone because it is the first time a software programmer attempted such a human-machine interaction with the goal of creating the illusion of human–human interaction.

ELIZA was brought together with another early artificial-intelligence program named PARRY for a computer-only conversation at the ICCC 1972 . While ELIZA is created to speak as a doctor, PARRY is created to simulate a patient with schizophrenia.[22]


Weizenbaum wrote ELIZA in MAD-SLIP for CTSS on an IBM 7094, as a software program to make natural-language conversation possible with a computer.[23] Weizenbaum identified five “fundamental technical problems” for ELIZA to overcome:

  1. the identification of critical words
  2. the discovery of a minimal context
  3. the choice of appropriate transformations
  4. the generation of responses appropriate to the transformation or in the absence of critical words
  5. the provision of an ending capacity for ELIZA scripts.[18] 

Weizenbaum solved the 5 problems and made ELIZA with no built-in contextual framework or universe of discourse[17] , requiring ELIZA to have a script of instructions on how to respond to inputs from users.[6]

AI Text Prompt for image ,using stable diffusion python script: “intricate schematic of ELIZA chatbots, many bots, many robots, chatterbots, super high deatils, 8k, hyper real, photorealistic, detailed –ar 16:9 –version 4 –s 42000 –uplight –no text, blur”

ELIZA starts its process of responding to human input by first looking at the text input for a “keyword”.[5] The “keyword” is a word defined by the human programmer as important by the ELIZA script that assigns a precedence number or a RANK to each keyword.[13] If keywords are found, they are put into a “keystack”, with the highest RANK keyword at the top. Then the input sentence is manipulated and transformed with the rule associated for the highest RANK keyword directions.[18] For example, when the DOCTOR script finds words “alike” or “same”, it outputs a message about similarity such as “In what way?”,[4] for the words with high precedence number. This also shows how certain words defined by the script, can be manipulated without contextual considerations, such as switching first-person and second-person pronouns and vice versa, as they both have high precedence numbers. Such keywords with high precedence numbers are defined as superior to conversational patterns and treated independent of contextual patterns.

After the first examination, the next step is to apply the appropriate transformation rule that includes: the “decomposition rule” and the “reassembly rule”.[18] The input is reviewed for syntactical patterns to establish the minimal context needed to respond. Using the input’s keywords and other nearby words, disassembly rules are tested until an relevant pattern is found. The sentence is then “dismantled” with the script rules and arranged into sections of the component parts with the decomposition rule for the highest-ranking keyword. A Weizenbaum example shows the input “You are very helpful”, which is transformed to “I are very helpful”. This is then broken into (1) empty (2) “I” (3) “are” (4) “very helpful”. The decomposition rule breaks the phrase into four small segments that contain the keywords and the information in the sentence.[18]

Then the decomposition rule designates a reassembly rule or set of reassembly rule swhen reconstructing the sentence.[5] The reassembly rule creates a response by taking the fragments of the input that the decomposition rule created, rearranges them, and adds in programmed words. In Weizenbaum’s previous example, the reassembly rule takes the fragments and applies them to the phrase “What makes you think I am (4)”, results in “What makes you think I am very helpful?”. This example is simple, but the output can be significantly more complex and use more of the input from the user. However, from this reassembly, ELIZA then sends the constructed sentence as text on the screen to the user.[18]

These steps show the bulk of the procedures that ELIZA follows to create a response from a typical input, but there are several specialized situations that ELIZA/DOCTOR can respond to. Weizenbaum wrote that when there is no keyword, to have ELIZA respond with a remark that lacked content, such as “I see” or “Please go on”.[18] Another method is to use a “MEMORY” structure that records prior recent inputs, and uses those inputs to create a response referencing a part of the earlier conversation when no keywords are found.[24] Slip’s ability to tag words for other usage makes it possible to simultaneously allow ELIZA to examine, store and repurpose words for usage in outputs.[18]

The exact manner that the ELIZA program dismantled, examined, and reassembled inputs is determined by the operating script. The script is not static and can be edited, or a new one created as needed for the operation in the context needed. This allowed the program to be applied in many situations, including the well-known DOCTOR script that simulates a Rogerian psychotherapist.[14]

Lisp version of ELIZA, based on Weizenbaum’s CACM paper, was written shortly after the paper’s publication, by Bernie Cosell.[25][26] A BASIC version appeared in Creative Computing in 1977 (although it was written in 1973 by Jeff Shrager).[27] Shrager’s version was ported to many of the earliest personal computers, appears to have been subsequently translated into many other versions in many other languages. Shrager claims not to have seen either Weizenbaum’s or Cosell’s versions.

Jeff Shrager searched MIT’s Weizenbaum archives in 2021, along with MIT archivist Myles Crowley and found files labeled Computer Conversations. The files included the complete source code listing of ELIZA in MAD-SLIP, with the DOCTOR script attached. The Weizenbaum estate gave permission to open-source the code under a Creative Commons CC0 public domain license. The code and other information can be found on the ELIZAGEN site.[26]

Another popular version of Eliza among software engineers is the version that comes with the default release of GNU Emacs, and which can be accessed by typing M-x doctor from most modern Emacs implementations.


From Figure 15.5, Chapter 15 of Speech and Language Processing (third edition). [28]

function ELIZA GENERATOR(user sentence) returns response
   Let w be the word in sentence that has the highest keyword rank
   if w exists
       Let r be the highest ranked rule for w that matches sentence
       response ← Apply the transform in r to sentence
       if w = 'my'
           future ← Apply a transformation from the ‘memory’ rule list to sentence
           Push future onto the memory queue
       else (no keyword applies)
               response ← Apply the transform for the NONE keyword to sentence
               response ← Pop the oldest response from the memory queue
   Return response


Weizenbaum was inspired by lay responses to ELIZA to write his book Computer Power and Human Reason: From Judgment to Calculation, where he explains the limits of computers and wants to make clear his opinion that anthropomorphic views of computers are just a reduction of the human being and any life form for that matter.[29] In the independent documentary film Plug & Pray (2010) Weizenbaum said that only people who misunderstood ELIZA called it a sensation.[30]

The Israeli poet David Avidan, who was fascinated with future technologies and their relation to art, wanted to explore the use of computers for writing literature. He conducted several conversations with an APL implementation of ELIZA and published them – in English, and in his own translation to Hebrew – under the title My Electronic Psychiatrist – Eight Authentic Talks with a Computer. In the foreword he presented it as a form of constrained writing.[31]

Many programs based on ELIZA are available in different programming languages. In 1980 a company called “Don’t Ask Software” created a version called “Abuse” for the Apple IIAtari, and Commodore 64 computers, which verbally abused the user based on the user’s input.[32] For MS-DOS computers, Sound Blaster cards came bundled with Dr. Sbaitso, which functions like the DOCTOR script. Other versions adapted ELIZA around a religious theme, such as ones featuring Jesus (both serious and comedic), and another Apple II variant called I Am Buddha. The a 1980 game The Prisoner used ELIZA-style interaction within its gameplay. In 1988 the British artist and friend of Weizenbaum Brian Reffin Smith created two art ELIZA-style programs written in BASIC, one called “Critic” and the other “Artist”, running on two separate Amiga 1000 computers and showed them at the exhibition “Salamandre” in the Musée du Berry, Bourges, France. The visitor was supposed to help them converse by typing in to “Artist” what “Critic” said, and vice versa. The secret was that the two programs were identical. GNU Emacs formerly had a psychoanalyze-pinhead command that simulates a session between ELIZA and Zippy the Pinhead.[33] The Zippyisms were removed due to copyright issues, but the DOCTOR program remains.

ELIZA is referenced in popular culture and continues to be a source of inspiration for programmers, developers, artists and humans focused on artificial intelligence. It was also featured in a 2012 exhibit at Harvard University titled “Go Ask A.L.I.C.E.“, as part of a celebration of mathematician Alan Turing‘s 100th birthday. The exhibit explores Turing’s lifelong fascination with the interaction between humans and computers, pointing to ELIZA as one of the earliest realizations of Turing’s ideas.[1]

ELIZA won a 2021 Legacy Peabody Award.


1969, George Lucas and Walter Murch incorporated an Eliza-like dialogue interface in their screenplay for the feature film THX-1138. Inhabitants of the underground future world of THX, when stressed, would retreat to “confession booths” and initiate a one-sided Eliza-formula conversation with a Jesus-faced computer who claimed to be “OMM”[citation needed]

1973, ELIZA influenced a number of early computer games by demonstrating additional kinds of interface designsDon Daglow wrote an enhanced version of the program called Ecala on a DEC PDP-10 minicomputer at Pomona College in 1973 before writing the computer role-playing game Dungeon (1975).[citation needed]

2011 video game Deus Ex: Human Revolution and the 2016 sequel Deus Ex: Mankind Divided features an artificial-intelligence Picus TV Network newsreader named Eliza Cassan.[34]

2016, Adam Curtis‘s documentary, HyperNormalisation, ELIZA was referenced in relationship to post-truth.[35]

January 2018, the twelfth episode of the American sitcom Young Sheldon, starred the protagonist “conversing” with ELIZA, hoping to resolve a domestic issue.[36]

August 12, 2019, independent game developer Zachtronics published a visual novel called Eliza, about an AI-based counseling service inspired by ELIZA.[37][38]


  1. Jump up to:a b “Alan Turing at 100”Harvard Gazette. 13 September 2012. Retrieved 2016-02-22.
  2. Jump up to:a b Berry, David M. (2018). “Weizenbaum, ELIZA and the End of Human Reason”. In Baranovska, Marianna; Höltgen, Stefan (eds.). Hello, I’m Eliza: Fünfzig Jahre Gespräche mit Computern [Hello, I’m Eliza: Fifty Years of Conversations with Computers] (in German) (1st ed.). Berlin: Projekt Verlag. pp. 53–70. ISBN 9783897334670.
  3. Jump up to:a b c d e f Weizenbaum, Joseph (1976). Computer Power and Human Reason: From Judgment to Calculation. New York: W. H. Freeman and Company. ISBN 0-7167-0464-1.
  4. Jump up to:a b Norvig, Peter (1992). Paradigms of Artificial Intelligence Programming. New York: Morgan Kaufmann Publishers. pp. 151–154. ISBN 1-55860-191-0.
  5. Jump up to:a b c Weizenbaum, Joseph (January 1966). “ELIZA–A Computer Program for the Study of Natural Language Communication Between Man and Machine” (PDF). Communications of the ACM9: 36–35. doi:10.1145/365153.365168S2CID 1896290 – via universelle-automation.
  6. Jump up to:a b c Baranovska, Marianna; Höltgen, Stefan, eds. (2018). Hello, I’m Eliza fünfzig Jahre Gespräche mit Computern (1st ed.). Bochum: Bochum Freiburg projektverlag. ISBN 978-3-89733-467-0OCLC 1080933718.
  7. ^ “ELIZAGEN – The Original ELIZA”sites.google.com. Retrieved 2021-05-31.
  8. Jump up to:a b Dillon, Sarah (2020-01-02). “The Eliza effect and its dangers: from demystification to gender critique”Journal for Cultural Research24 (1): 1–15. doi:10.1080/14797585.2020.1754642ISSN 1479-7585.
  9. ^ Bassett, Caroline (2019). “The computational therapeutic: exploring Weizenbaum’s ELIZA as a history of the present”AI & Society34 (4): 803–812. doi:10.1007/s00146-018-0825-9.
  10. ^ “The Samantha Test”The New Yorker. Retrieved 2019-05-25.
  11. ^ Marino, Mark (2006). Chatbot: The Gender and Race Performativity of Conversational Agents. University of California.
  12. ^ Colby, Kenneth Mark; Watt, James B.; Gilbert, John P. (1966). “A Computer Method of Psychotherapy”. The Journal of Nervous and Mental Disease142 (2): 148–52. doi:10.1097/00005053-196602000-00005PMID 5936301S2CID 36947398.
  13. Jump up to:a b Shah, Huma; Warwick, Kevin; Vallverdú, Jordi; Wu, Defeng (2016). “Can machines talk? Comparison of Eliza with modern dialogue systems” (PDF). Computers in Human Behavior58: 278–95. doi:10.1016/j.chb.2016.01.004.
  14. Jump up to:a b Shrager, Jeff; Berry, David M.; Hay, Anthony; Millican, Peter (2022). “Finding ELIZA – Rediscovering Weizenbaum’s Source Code, Comments and Faksimiles”. In Baranovska, Marianna; Höltgen, Stefan (eds.). Hello, I’m Eliza: Fünfzig Jahre Gespräche mit Computern (2nd ed.). Berlin: Projekt Verlag. pp. 247–248.
  15. ^ Weizenbaum 1976, p. 188.
  16. ^ Epstein, J.; Klinkenberg, W. D. (2001). “From Eliza to Internet: A brief history of computerized assessment”. Computers in Human Behavior17 (3): 295–314. doi:10.1016/S0747-5632(01)00004-8.
  17. Jump up to:a b Wortzel, Adrianne (2007). “ELIZA REDUX: A Mutable Iteration”. Leonardo40 (1): 31–6. doi:10.1162/leon.2007.40.1.31JSTOR 20206337S2CID 57565169.
  18. Jump up to:a b c d e f g h i Weizenbaum, Joseph (1966). “ELIZA—a computer program for the study of natural language communication between man and machine”. Communications of the ACM9: 36–45. doi:10.1145/365153.365168S2CID 1896290.
  19. ^ Wardrip-Fruin, Noah (2009). Expressive Processing: Digital Fictions, Computer Games, and Software Studies. Cambridge, Massachusetts: MIT Press. p. 33. ISBN 9780262013437OCLC 827013290.
  20. ^ Markoff, John (2008-03-13), “Joseph Weizenbaum, Famed Programmer, Is Dead at 85”The New York Times, retrieved 2009-01-07.
  21. ^ Weizenbaum, Joseph (1976). Computer power and human reason: from judgment to calculation. W. H. Freeman. p. 7.
  22. ^ Megan, Garber (Jun 9, 2014). “When PARRY Met ELIZA: A Ridiculous Chatbot Conversation From 1972”The AtlanticArchived from the original on 2017-01-18. Retrieved 19 January 2017.
  23. ^ Walden, David; Van Vleck, Tom, eds. (2011). “Compatible Time-Sharing System (1961-1973): Fiftieth Anniversary Commemorative Overview” (PDF). IEEE Computer Society. Retrieved February 20, 2022. Joe Wiezenbaum’s most famous CTSS project was ELIZA
  24. ^ Wardip-Fruin, Noah (2014). Expressive Processing: Digital Fictions, Computer Games, and Software Studies. Cambridge: The MIT Press. p. 33. ISBN 9780262013437 – via eBook Collection (EBSCOhost).
  25. ^ “Coders at Work: Bernie Cosell”codersatwork.com.
  26. Jump up to:a b “elizagen.org”elizagen.org.
  27. ^ Big Computer Games: Eliza – Your own psychotherapist at http://www.atariarchives.org.
  28. ^ https://web.stanford.edu/~jurafsky/slp3/15.pdf
  29. ^ Berry, David M. (2014). Critical theory and the digital. London: Bloomsbury Publishing. ISBN 978-1-4411-1830-1OCLC 868488916.
  30. ^ maschafilm. “Content: Plug & Pray Film – Artificial Intelligence – Robots”plugandpray-film.de.
  31. ^ Avidan, David (2010), Collected Poems, vol. 3, Jerusalem: Hakibbutz Hameuchad, OCLC 804664009.
  32. ^ Davidson, Steve (January 1983). “Abuse”Electronic Games. Vol. 1, no. 11..
  33. ^ “lol:> psychoanalyze-pinhead”IBM.
  34. ^ Tassi, Paul. “‘Deus Ex: Mankind Divided’s Ending Is Disappointing In A Different Way”Forbes. Retrieved 2020-04-04.
  35. ^ “The Quietus | Opinion | Black Sky Thinking | HyperNormalisation: Is Adam Curtis, Like Trump, Just A Master Manipulator?”The Quietus. Retrieved 26 June 2021.
  36. ^ McCarthy, Tyler (2018-01-18). “Young Sheldon Episode 12 recap: The family’s first computer almost tears it apart”Fox News. Retrieved 2018-01-24.
  37. ^ O’Connor, Alice (2019-08-01). “The next Zachtronics game is Eliza, a visual novel about AI”Rock Paper Shotgun. Retrieved 2019-08-01.
  38. ^ Machkovech, Sam (August 12, 2019). “Eliza review: Startup culture meets sci-fi in a touching, fascinating tale”Ars Technica. Retrieved August 12, 2019.



%d bloggers like this: