“Men will set the goals, formulate the hypotheses, determine the criteria, and perform the evaluations.”
In his 1960 paper “Man-Computer Symbiosis,” J.C.R. Licklider imagined a partnership in which humans and computers would contribute different strengths. He wrote, “Men will set the goals, formulate the hypotheses, determine the criteria, and perform the evaluations.” The language reflects its era, but the allocation of responsibility remains strikingly current. People would decide what mattered and how success should be judged; machines would handle routine operations, search, calculation, and the rapid comparison of alternatives. Licklider was not describing a passive tool or an autonomous replacement. He was outlining a cooperative system designed to expand the reach of human inquiry.
The idea challenged the dominant batch-computing model of its time. Instead of preparing a job, handing it to a distant machine, and waiting for output, Licklider envisioned interactive computing: a person and a computer exchanging information quickly enough to shape each other’s next step. That vision helped inspire personal computing, networked communication, and interfaces that make computation part of an ongoing conversation. Modern AI assistants look like another turn of the same ambition. They can summarize, generate options, analyze patterns, and respond in natural language, reducing the friction between an intention and a computational operation.
Yet the quote places a boundary around that convenience. Goals and criteria cannot be treated as neutral details. A model optimized for engagement may spread sensational material; one optimized for efficiency may ignore fairness or dignity. Even evaluation is value-laden: accuracy averages can conceal failures for smaller groups, and productivity measures can omit quality, autonomy, or trust. Licklider’s division of labor asks organizations to state what they want before they automate pursuit of it. It also asks them to keep people capable of questioning whether the original goal was wise when conditions change or unexpected effects appear.
Human-computer symbiosis is most valuable when it creates a feedback loop without dissolving accountability. A clinician may use AI to surface evidence while retaining responsibility for diagnosis. A journalist may use it to scan documents while verifying claims and protecting sources. A scientist may generate hypotheses while deciding which deserve costly experiments. In each case, the computer broadens the field of possibilities and the human supplies purpose, context, and judgment. Licklider’s sentence is therefore a design principle for today: build systems that make human reasoning more informed and capable, while preserving the authority to set aims, evaluate results, and change course. The partnership should leave people better able to understand and govern the work, not merely faster at accepting machine suggestions. That standard turns symbiosis from a metaphor into a measurable design goal: stronger human agency supported by computation that remains visible, contestable, and subordinate to chosen purposes.
Licklider proposed a close, interactive partnership rather than a computer that merely completed isolated batches of calculations.
He expected machines to accelerate search and routine work while people supplied direction, hypotheses, and judgment.
Optimization only works after someone decides what outcome matters and how it will be measured.
Those decisions can encode social priorities, omissions, and tradeoffs, so they should remain visible and open to challenge.
AI can help clinicians review evidence, scientists explore hypotheses, and journalists examine large collections of records.
The partnership is strongest when expertise is amplified without handing final responsibility to an opaque system.
Explore more "Quotes of The Day"
Continue exploring how writers, thinkers, artists, and advocates have described the choices that shape an everyday life.