Is Anthropomorphic Design a Viable Way of Enhancing Interface Usability?

Chapter Five: Conclusion – Is Anthropomorphic Design a Viable Way of Enhancing Interface Usability?

After examining the results of the questionnaire and relevant literature on anthropomorphism it seems indisputable that a significant proportion of people do anthropomorphise their computers and regard them as social actors. Should we embrace this natural phenomenon and develop systems which support it as a method of interaction? Or should we attempt to avoid it, for fear of where it may lead?

The test results in Chapter Four indicate that users did prefer using the anthropomorphic interface, but does this mean that it is better? Further investigation is required to ascertain whether users would still enjoy this system after a prolonged period of use – would it ultimately become annoying? – and is enjoyment the most appropriate measure of success? Different designers have different ideas on how the success of a humanlike interface should be determined.

One of Shneiderman’s concerns is that although users may well enjoy using a humanlike interface, controlled studies have failed to prove that they enhance productivity (personal communication, 2005)

Brenda Laurel suggests that anthropomorphism works ‘more or less’ well, depending on the context of use, citing the ‘Guides’ Project at Apple as a good example of a useful tool, but condemning the idea of any help agent that pops up as ‘patently absurd’ (personal communication, 2005).

Where the context of use is one that does not depend on high productivity, anthropomorphism may be appropriate as it could improve the user experience in other ways. In a learning environment it has been proven to improve task persistence and self efficacy (Fogg & Nass, 1997). As mentioned in Chapter Two, the Department of Communication at Stanford University have recently been examining the influence of a ‘Colearner’ anthropomorphic agent in the context of an interactive learning environment, with positive results (Ju, Nickell, Eng & Nass, 2005).

If current trends in customisation continue then the ‘celebrity interface’ described in Chapter Four could be applicable as a means of making interaction more enjoyable and reducing stress. As previously noted, it would only need to provide minimal cues in order to provoke a social response. The limitations of current technology should be observed in the design, it should be implemented in a subtle way and should not aspire to unrealistic capabilities. It would also need to be applied in the right context. It is noted in the Colearner study (ibid.) that the “casting of personality” is very important. Tthe diversity of choice displayed by users in Chapter Three should be considered – along with the ability to turn it off with ease, for those who choose not to interact in this way.

Anthropomorphic characters have often been included in science fiction because they make interesting plot devices and often feature as characters in their own right. The idea of direct manipulation may sound less exciting, but in terms of future technology, it is infinitely more viable. Consider this far-out vision of the future of direct manipulation, from an interview with Terence McKenna (Hodgkinson, 1993):

“As technology becomes ever more pervasive, it should become ever less visible…here’s one possible scenario…[where] technology has been reduced essentially to a pair of black contact lenses that are surgically planted…onto the under surface of the eyelid, so when you close your eyes, there are menus hanging in space…computer menus. And by squinting or blinking you open menus, and the entire culture could be virtualised in this way.

So appearing to be primitive, we would actually have access to enormous databases, all kinds of situations of education, entertainment, personal interaction – but there would be no visible manifestation of technological involvement other than these contact lenses, which could run off ambient static electricity and would be manufactured in a factory on the moon, or something. In other words, a complete abandonment of physical technology, but a complete maintenance of the capacity of physical technology.”

If the controls were in our eyelids, would we still consider the computer to be a social actor? The anthropomorphic interface would be obsolete. Why ask ‘someone else’ (ie an anthropomorphic character) to do it, when it is quicker to do it yourself by blinking your eyes? Perhaps this is why Shneiderman becomes so animated when you mention anthropomorphism. His vision is beyond the need for humanlike interaction, it is direct. At present we interact with our computers socially because they are difficult to use and there is still very much a ‘them and us’ (human vs. machine) scenario. We apply social rules to negate feelings of fear or confusion and to provide our own context within which to comprehend the complexity of the technology. It is a credit to the human race that we have been able to develop such inventive ‘work-arounds’ when the technology repeatedly fails us..

If anthropomorphism is inevitable, we should embrace it for now. Whilst the technology is still immature, it may improve the user experience to acknowledge these beliefs and work with them, in appropriate contexts. As our systems expertise continues to grow, the technology should become less visible and our interactions more transparent – then there will be no need for us to chat to our computer. We might designate our most mundane tasks to an autonomous agent, but increased usability will ultimately reduce the amount of time we spend grappling with inferior systems and give us more free time to get out and interact socially with real humans.

© Alison Flind 2006