I like the MRIAB concept.
Been working on this myself.This Market Researcher in a Box could sift through all that juicy data on the nets, figure out trends, decide out what consumer products are “needed” feed them to the EIAB which then designs them.
The problem is that with the same market research, everyone comes to the same conclusions. A friend of mine was just complaining to me the other day about that concerning cars. He says all cars look alike lately and have very similar lines. Because market research says "cars with squarish windows are selling well lately". So every car company produces car with squarish windows. I don't know how much truth there is to that since i'm not really a car person, but i do see how it could be a problem in the future. Trends don't account for new information, they are based on the past. Without PEOPLE giving us products that we didn't know we wanted and being creative, I fear it will all lead to a sad homogeneity.
Shriekback loops are a major problem with flesh and blood researchers but maybe our MRIAB could use some pseudo-randomization to shake things up remember the low overhead reducing risks that come from a greater heterogeneity of products. the MRIAB as to be sophisticated perhaps figuring out that a car with a oval window would stand out in a crowd and taking advantage of that. Maybe there would be work for human product testers and focus groups until the MRIAB figures out a working model of consumption. (for full disclosure I think this future would suck as well)
very interesting what problems have you run into? What do you think the limits of AI are in market research?
What's so interesting to me is how easy it is to tap into all the available feedback looks. Just look at kickstarter or indiegogo. You essentially can throw ideas / concepts / prototypes out to the masses to fund, thus confirming the market's viability. It's possible to create concepts and throw them out to the market and see what sticks. Alternatively, you can do the opposite. Listen for market demands and once identified, fulfill those needs. There is a treasure trove of data on social media that can show sentiment and needs in the aggregate. Exciting stuff. If the aggregate isn't your bag, just mine individual needs and create one off solutions for those individuals... your options are almost limitless. The biggest challenge is making your artificial research appear human... or put another way not spamming feedback providers.To get something that hits, you need scale and that requires the appearance of a human product launch.
Ultimately it I don't think we would have to pretend it was human.
One thing that I don't think this thought experiment considers is that Apple created a market. No one thought that computers would have value to individual consumers. It took a leap of faith on Jobs' part to design, build, and market something that no one even wanted. I have trouble seeing how a machine or algorithm can ever accomplish something like that.
Apple was not the first Personal computer or even in the first generation of personal computers so Steve did not really create anything other than a slicker version of something that already existed. Steve Jobs is more of a Don Draper than a Thomas Edison.
I'll agree with you about Jobs. I used him and Apple as an example because they're easier to grasp and everyone know about them. I still think the point holds though; how can a computer or algorithm know what people want, before they want it? How can machines take a leap of faith?
faith is treating things you don't know like you do. Computers are great at that.