MIT Artificial Intelligence Laboratory

A laboratory of its own

The MIT Artificial Intelligence Laboratory was formed in 1970, when many of Marvin Minsky's AI colleagues left Project MAC to join him in the new laboratory^[1]. Minsky heads the list of the laboratory's directors, serving from 1970 to 1972^[1].

The date marks an administrative separation rather than the beginning of AI work at MIT. Research on artificial intelligence at MIT began in the late 1950s, a separate matter from the laboratory's 1970 founding^[1], and the MIT Computer Science and Artificial Intelligence Laboratory (CSAIL) dates the AI project itself to 1959^[2].

Out of Project MAC

In the early 1960s an "AI Group" was incorporated into Project MAC; its director was Minsky, and it included John McCarthy, the inventor of Lisp^[1]. Project MAC was launched on 1 July 1963 with a $2 million grant from the Advanced Research Projects Agency^[1]. Its first director was Robert Fano, listed as serving from 1963 to 1968, while J. C. R. Licklider was the DARPA programme manager^[1]. CSAIL's own history page describes Project MAC as a project sponsored by the Department of Defense to develop a computer system accessible to a large number of people^[2] — that is, a time-sharing system, in which one machine's processor time is divided among many users instead of being given to one person at a time.

Space, and the split

By the late 1960s Minsky's AI group sought more space and could not get it from Project MAC's director, Licklider^[1]. Minsky found that while Project MAC as a single entity could not get the additional space, splitting off to form his own laboratory would entitle him to more office space^[1]. On that account the new laboratory began as a way of securing room to work, at a moment when computing at MIT was expanding.

Systems and machines

Through the 1960s and 1970s the AI Group developed the Incompatible Timesharing System, which ran on PDP-6 and later PDP-10 computers^[1]. A time-sharing system of this kind lets many programmers use one machine at the same time, each from a terminal, which is what makes a shared laboratory style of research practical.

The laboratory also invented Lisp machines — computers built around running Lisp, the language McCarthy had designed — which were commercialised by Symbolics and Lisp Machines Inc. in the 1980s^[1]. A Lisp machine put the language into the hardware, so that programs written in it ran fast enough for interactive work.

Into CSAIL

Project MAC was officially renamed the Laboratory for Computer Science in 1976^[1]. As collaboration across the two laboratories increased, they merged on 1 July 2003 to form the MIT Computer Science and Artificial Intelligence Laboratory, CSAIL^[1][2]. CSAIL's history page describes the new laboratory as having its roots in two MIT computing powerhouses, the Laboratory for Computer Science and the Artificial Intelligence Laboratory^[2], and reports more than 60 research groups working on hundreds of diverse projects^[2].

Place in the story

The AI Lab existed as an institution in its own right from 1970 until the 2003 merger. Among the things built inside it were a time-sharing operating system for the PDP-6 and PDP-10 line and the Lisp machine, which became a commercial product through Symbolics and Lisp Machines Inc. in the 1980s^[1]. Its history also shows how closely the organisation of research and the funding behind it were tied together: a group inside a defence-sponsored project on shared computing, a separate laboratory created partly to obtain office space, and finally a merger with the renamed Project MAC once the two sides were already collaborating^[1][2]. The three dates — AI work from the late 1950s, the laboratory in 1970, the merger in 2003 — are usually run together in retellings, but they mark different things.