Design humans in or design humans out?

In the most recent issue of ACM's Ubiquity, Francis Hsu argues for "lowering the frequency and necessity of human data inputs" in future IT systems and applications.

When I worked at Pacific Bell many years ago, I often heard similar arguments. IT systems were designed to minimize human involvement in their operation. Humans were to be involved only when there were "exceptions" -- i.e., cases that the technology could not handle on its own. The goal was to reduce this pricey human involvement as much as possible.

Reducing pricey human involvement remains the goal years later for lots of systems. The Vice President responsible for usability and user productivity at a major enterprise software and services company emphasized the importance of this goal in a conversation I had with him earlier this year.

Contrast the above perspective with that of John Thackara, who visited the San Francisco Bay Area in May to promote his book "In the Bubble: Designing in a Complex World." John lamented the ongoing goal of replacing people with technology, telling tales about the terrible user experiences that so often result. Referencing very different, innovative examples exhibiting outstanding user experiences, John advocated a design principle of "enabling human agency" -- of designing people in, rather than designing people out.

Should humans be designed in or designed out? Is the answer, "it depends"? If so, on what does it depend?

For more information on John's new book, including extracts from the book, see www.thackara.com/inthebubble/.