This looks like an interesting project.
Not sure if I would call it AI either. Looks like a basic chatbot with a avatar on top.
On the other hand, as John McCarthy says, if it works, it’s not AI.
I agree these chat bots seem a bit clunky in it’s current form but to be able to project forward a few years and to what types of data we will be able make sense of, and how they relate to someone’s emotional state will have a profound effect on how effective these virtual humans are. SM already uses data points from the face and tone of voice to assess mood. I think it’s going to surprise a lot of people how indistinguishable virtual humans will be from real therapists in a few years
.
I’d be very surprised indeed, and very concerned for that matter. I also do not believe that is the goal at all: I think virtual people will be very much discernable from real humans, and that they will assist in certain tasks (like providing health information, offering a 24/7 service, or interpreting self-tracking data and giving personalised advice based on this), but will not be replacing actual therapists due to the sheer complexity of that interaction. Much like a virtual reality environment will not be indistinguishable from real life, nor should we want it to be, but it may be great for training purposes in a healthcare setting.
I agree, Tania and we have been here before with the AI therapist in earlier iterations of AI. The point you make was true for exactly the reasons you say. Therapy is complex diverse and idiosyncratic.
I have often wondered why the debate has not focused on managing large-scale but lower-level problems like loneliness. In what ways is an AI chatbot in your smart speaker better than the cat?" seems a good research project for someone.
True Greig, and I do have this type of discussion with my students in the Digital Health Ethics course because these are important topics to tackle as we progress in the development of these technologies. Lots of underexplored ethical challenges, like deception, technological resistance, function creep etc, but also some fascinating work being done e.g. with robot pets, remote monitoring technology and smart homes.