To most of the public, AI went from being a far-fetched concept of science fiction to something intricately linked to our work, entertainment and media. However, the history of AI stretches back much further than the 21st century. Since its conceptual beginnings in the 1950s, AI technology has been seen by the public with fear, doubt and optimism, influenced by media portrayals from dystopian films like 2001: A Space Odyssey (1968), and The Terminator (1984), as well as prominent voices in the scientific community like Allen Newell and Herbert Simon in the 1950s and 60s who predicted that AI would beat humans at chess and mathematics in a decade.
British logician Alan Turing first posed the question of whether machines could think, written in his paper “Computing Machinery and Intelligence,” published in 1950. He wrote this paper just a few years after the first digital computers were created, and established what became known as the “Turing test” to determine whether a machine could think. He stated that a machine could be considered intelligent if a person conversing with it could not tell whether they were talking to a human or a machine.
Two years after Turing’s death, the founding moment of AI research occurred in 1956, when John McCarthy, a professor at Dartmouth College, organized the Dartmouth conference to discuss “how to make machines use language, form abstractions and concepts, solve kinds of problems now reserved for humans and improve themselves.” The focus became developing AI that could simulate human intelligence at expert-level tasks, such as playing games like checkers and chess, language processing and mathematics.
A breakthrough came when a program called SAINT (symbolic automatic integrator) was developed by James Slagel in 1961, which could solve calculus problems at the level of a college freshman. It was one of the first ever “expert systems” – computer systems which could solve problems at the level of a human expert.
Then, in 1966, ELIZA, the first-ever chatbot was created. ELIZA scanned users’ input for keywords, and based on them selected a pre-programmed response incorporating the keyword, usually an open-ended question. It made conversations based on speech patterns rather than true comprehension, giving the illusion of intelligence; the chatbot was incapable of understanding the content of its conversations. While it followed simple rules for its responses, it was a step up from the list of prewritten responses that earlier models used. It was modeled after famous psychotherapist Carl Rogers, mimicking his open-ended style of therapy sessions.
However, Joseph Weizenbaum, the creator of ELIZA, was skeptical about the future capabilities of his chatbot and ones like it. He believed the chatbot was only a tool, and that the program was fundamentally limited in its ability to understand human emotions and intellect.
In the mid-60s, AI research was heavily funded by the Department of Defense, but after the initial boom of interest in the 1960s, government funding for AI research started to dry up in the 1970s. This was largely due to the overhype surrounding AI research and under delivery due to the limited computing power available at the time. The first of these “AI winters,” as these periods of halted progress were called, was largely set in motion by the ALPAC (Automatic Language Processing Advisory Committee) report of 1966, which found that despite millions of dollars in funding, automatic translations were no better than human translators, and much costlier.
Other prominent critics included Sir James Lighthill, a celebrated British mathematician who wrote the “Lighthill Report” in 1973, primarily responsible for shutting down AI research in the U.K. He expressed his skepticism of the potential of AI to solve complex, real-world problems.
In the 1990s and 2000s, with large improvements in computer speed and data processing capabilities, AI research saw a comeback, shifting from the rule-based programs of the past to data-driven systems. With the release of OpenAI’s ChatGPT in November of 2022, AI moved from a niche tool to mainstream application. Conversational, generative AI changed the way everyday people worked, and brought with it an array of concerns.
With significant advancements in AI technology, modern concerns over AI range from ethical issues to AI becoming advanced enough to take over jobs. 50% of U.S. adults feel more concerned than excited about the future impacts of AI, and 53% believe it will worsen people’s ability to think creatively. A MIT Media Lab study suggests that using LLMs – large language models – in the writing process could harm brain activity, bypassing memory storage and active engagement. However, the study also suggests that if used properly, AI could enhance learning rather than inhibiting it. Overall, the extent of the future impact and growth of AI technology remains to be seen.
