There are a number of issues coming from this episode which are very interesting to talk about. I'd like to focus briefly on one that really caught my ear.
A lot of analysis centred on how collecting information about our browsing and purchasing habits provides a business with information to tailor its services ever more effectively. Reed Hastings, CEO & Founder of Netflix, the DVD rental outfit had this to say:
We look at movies as a really rich area to try to understand human behaviour and how to create a better experience than any other video system so that people watch more and more movies.The show's presenter Dr Krotoski immediately sounds a note of caution:
Movie taste is very personalised. But what we realised is is if we ask people to tell us what they've loved in the past that our computer systems can do a really good job of helping them choose movies that they're more likely to enjoy in the future.
I worry that in the process we've lost something. I wonder if recommendation systems don't defeat the point of the web. Isn't the vast possibility that the web offers for serendipity to bring us unexpected new ideas from accidental encounters being replaced by a process that apparently broadens our horizons but actually sells us the same things.In terms of the web as a big marketplace, I think its easy enough to argue either side of this effectively and, I might add, maintaining the integrity of one's intentions.
Yet there's a more abstracted point to be made here, a sort of meta-view away from the hard information on the screen. Another sceptic on the programme, Douglas Rushkoff, puts this succinctly, if with cold irony:
Recommendation engines are very good at figuring out what people like me would do and telling me what that is so I can then find out what people like me do. I can become much more like a person like me.When we take it upon ourselves to encounter art we might choose something as it's familiar, or something else precisely because it's not. Either way, I think there's an important critical faculty that one needs to employ at all times. This is an extension of the moral judgement that also forms the issue of aesthetics. To allow the market that has presented us with a previous artistic experience to then narrow the choices open to us by prescribing further encounters, takes that critical faculty away from us incrementally.
Reed Hastings goes on to say:
We are 100% about trying to improve our consumers' enjoyment of movies. We help them get the movies that they're going to laugh at most, cry the most [sic], love the most.This is the inevitable consequence of the situation, that an assumption is made whereby a) the artistic event, or moment, that causes laughter, crying and so on can be quantified b) that increasing that quantity delivers more enjoyment to the viewer. As an algorithmic model this is sound logic but in aesthetic terms utter nonsense. Above all it ignores the possibility that being able to make our own critical judgement about a film is that which constitutes our "enjoyment" in the first instance.
Hastings does end with:
It's all about pleasing the consumer and if that narrows that's fine, if that broadens that's fine.although this reasonable comment does seem to contradict the trend-identifying process he was actually trying to sell.