Marketing is imitation
(do large, data-driven AB tests, find what works, introduce incremental change, improve. Repeat ad infinitum. Imitation of existing models combined with data analysis provides ‘solutions’ that are likely in an overwhelming majority of cases to work.
Test cases for charity include:
- pictures of children
- presence of an ask
- freepost return envelope
- presence of a signature
- use of certain colours
- slightly sad faces
- urgent situation that needs addressing)
Many of the uses of AI operate using the same key assumptions.
AI can gather and analyse masses of data almost instantaneously, and return a solution based on an algorithm. It can thus provide medical diagnoses (based on massive and growing libraries of case studies), throw up legal issues for companies (based on millions of contracts), marketing recommendations for books/ music/ films to enjoy (based on the selections made by you and millions of others) etc.
AI can do imitation; it can do incremental learning. Nascent AI can do much more – come up with new, creative and entirely unexpected solutions to problems (e.g. in highly complex games like chess and Go!) Combine this with massive data sets on consumer entertainment preferences, and AI-designed houses tailored to the individual, then AI-authored films, music, novels become entirely foreseeable.
(Gallai AI, gyda chorpysau Cymraeg digon o faint, ac o ddysgu rheolau’r gynghanedd, ysgrifennu awdlau ac englynion hyd yn oed, yn y pendraw.)
But analogy. Humans can draw analogies across disciplines, fields, diverse experiences and use this ability to create entirely novel things. AI can’t use sex/ fine wine/ dreams/ the shape of clouds to bring an insight to a scientific/ engineering/ artistic conundrum.
This would imply an important ability that humans have which AI would lack, at least until such a point was reached where AI was able to imitate so many aspects of human behaviour in so many spheres that it was starting to meaningfully imitate humans, in the round.
But then humans remain the measure of AI, the telos to which AI is directed, in the real world simply because ‘the economy’, by definition is centred around and measures value by human needs and desires.
Key questions:
- Are human fashion trends primarily consumer-driven, or producer/marketer-driven? [Cf. Natural language change, as a large and universal test-ground for status-driven change amongst humans, implies that to a great extent they are at a deep level human driven, and thus unpredictable.]
- Is there sufficient available energy (and finite precious earth resources) to sustain an AI supercomputer-driven economy? https://www.newscientist.com/article/2205779-creating-an-ai-can-be-five-times-worse-for-the-planet-than-a-car/ If not, can a microcosm global economy exist that is highly resource-intensive? (i.e. one that does not include all the earth’s population or the entire earth geographically, but can maintain itself despite this?)
- The ergonomic issue. People prefer a world made by hand, when offered the choice. So an internet-driven economy is not ergonomic, simply because it meaningfully engages at most two of the five senses – sight and sound, and ignores the rest of the body (we are psycho-somatic, emotional, spiritual and fully embodied beings). What implications do these realities have for an AI-shaped economy, and its limitations?