Page 195 - Artificial Intelligence for the Internet of Everything

Would IOET Make Economics More Behavioral? 181

own attention among four classes of activities: listening, storing, thinking
and speaking. A general design principle can be put as follows: An informa-
tion processing subsystem (a computer or new organization unit) will reduce
the net demand on the rest of the organization’s attention only if it absorbs
more information previously received by others than it produces—that is, if
it listens and thinks more than it speaks.” (Ibid, p. 42; italics added). Briefly put,
IoE has to make us feel quieter rather than noisier.
So far there is no clear evidence to show that our IoE environment can
satisfy the Simon condition, and it is not entirely implausible to say that it tends
to speak more than it can effectively listen and think (Chen, Chie, & Tai,
2016; Chen & Venkatachalam, 2017). If so, while a myriad of interconnec-
tions and interactions provided by IoE allow us to surf over a huge space of
opportunities, it also exposes us to a potentially large number of decision
problems, each with many alternatives. The latter is notoriously known
as the choice overload problem or the paradox of choice (Iyengar & Lepper,
2000; Schwartz, 2004). Time and attention allowed for each of these choice
problems is, therefore, severely diluted. Under such circumstances, to facil-
itate decision making, the attention-lacking agents may rely more on their
fast track of information processing (i.e., the reflexive system), and less on their
slow or deliberate track, the reflective system (Kahneman, 2011). In addition to
emotion and gut feeling, various fast and frugal heuristics, such as following the
herd, choice reinforcement, or using rules of thumb, will play a more con-
tributory role in decision making (Gigerenzer & Gaissmaier, 2011), which
may again make decision makers more like homo sapiens instead of homo
economicus.
Finally, if information overload and choice overload have driven deci-
sion makers to behave more like homo sapiens, then even though machine
learning can effectively extract and learn the behavioral patterns of these
decision makers, the artificial agents that have been built may be, at best,
another homo sapiens, since what was learned by artificial agents is what they
actually did, but not what they ought to do for the sake of their own best
interest. If one employs these artificial agents as the incarnation of their
human counterparts and automates the decisions for them, then the
well-known GIGO (garbage in, garbage out) principle may be applied
(Stephens-Davidowitz & Pabon, 2017), and the things that are in action
are again homo sapiens and not homo economicus.
The above three cases, while not exhaustive, justify why Thaler’s predic-
tion remains valid, and is independent of the IoE technology.