I recently had the pleasure of attending a small workshop on Trust and Cloud Computing sponsored by Corpus Christi college at Cambridge University (UK) and Microsoft’s Cambridge Research group. It was designed to bring philosophers, technologists, sociologists, and other academic stripes under one roof to focus their efforts on trying to understand some of the trust implications of cloud computing.
The workshop, exceptionally organized by Tom Simpson, proved to be thoroughly interesting and fruitful. There is a great synopsis of the workshop here, which was written by one of the other attendees. It’s well worth a read.
Missing from the synopsis is an overview of Ian Kerr‘s public keynote address, titled “in machines we trust? cloud computing, ambient intelligence and robotics.” Dr. Kerr, who holds a Canada Research Chair in Ethics, Law, and Technology at the University of Ottawa, argued that we ought to be wary of trusting machines (trusting them is different from relying on them to do things). The field of affective computing–a goal of which is to foster emotional attachment between humans and robots through the design of “emotional robots”–is advancing to the point that artificial software agents, “bots” and the like, are currently able to foster those relationships. “Chatter bots” engage people in online chats oftentimes to solicit personal information from those individuals. To the unsuspecting chat participant the underlying motivation for those relationships appears to be benign–the bots are just making conversation, and allowing their conversational partners an opportunity to engage in harmless chatter. However, as Dr. Kerr points out, the real motivation for fostering online relationships is to satisfy a business model that seeks to gather personal data, often for marketing purposes, a fact that is obfuscated to the human conversational partner.
In the parlance of philosophers, bots in the cloud might be trusted, but the disjoint between their outward facade and the underlying corporate motivations makes them untrustworthy to the core.
The kinds of technical capabilities represented by cloud computing, for both data mining and natural language comprehension (the kind of comprehension required for meaningful conversation), far surpass those available today.
The workshop was a fantastic start to what will hopefully become a collaboration between academics who have an interest in keeping the cloud trustworthy in the philosophical sense of the term.