Implications of Human Heuristics as the Analogue(s)
for an Input Decision Tree for Autonomous Agents
Heuristics (Greek meaning to
discover) is said to involve the exploration of problem-solving techniques
which serve to improve performance by self-education. It can be
narrowly debated that this process is best described as a methodology
or a type of logic for discovery using already existing rules, while
heuristics are the rules of discovery. For our purposes, the components
that allow and set the parameters of human heuristics or discovery will
be examined. An infinite being is omniscient and has nothing to
discover but finite beings with self-awareness, humans, can and do heuristically
The heuristic structure can
be seen as an analogue (meaning to be similar in structure yet different
as in an analogy) for the human creations of robots and other machinery.
Those creations are called autonomous agents and will be discussed in
the next section of this paper. Autonomous robots and machinery
can have the input modeled from a basic human decision tree.
The accuracy of complex human
decisions has been questioned by many researchers. Philip E. Tetlock's
2005 book, titled Expert Political Judgment, reported on a two-decade
long study of predictions by experts. He had 284 subjects that
ranged in status from expert to non-expert. Tetlock collected
82,361 forecasts to be evaluated from his experts, informed non-experts,
and uninformed non-experts. The experts barely, if at all, outperformed
the informed non-experts. That result was not new. In 1980, J.
Scott Armstrong suggested that no matter how much evidence exists that
seers do not exist, suckers will pay for the existence of seers.
Jerome Groopman wrote in his
recent book, How Doctors Think, that studies indicate as many
as 15 percent of hospital patients receive inaccurate diagnoses. Doctors
are not generally taught heuristics in medical school. However,
doctors are coming to reason from flow charts and algorithms much like
computers. This may prevent diagnoses largely based on how often
a symptom has been seen or diagnoses made from reluctance to pursue
less likely diagnoses from fear of being criticized if inaccurate.
Obviously, as mankind increases
epistemologically so will the potential for better performance when
applied to his creations.
2. Autonomous Agents
Autonomy is a Greek word referring
to a state where one gives themselves their own law or is self-governing.
Humans are self-aware and have freewill to choose a course of action
from various alternatives. The human agent has autonomy in pursuing
individual needs within compliance or noncompliance of behavioral parameters
or natural law. Those natural laws are widely known as the Decalogue
and its subsets. The individual choice of theology will determine
how those parameters are accepted or rejected as will be the characteristics
of personality, such as love and justice.
Locke saw natural law as epistemologically
empirical. The reality of intangible natural law is empirically
evident in the efficiency of behavioral compliance with the laws and
the inefficiency of behavioral noncompliance with the laws. (Spurious
objections to this truth have been analyzed in earlier papers in this
series.) Kant argued that an autonomous agent is demonstrated
by decisions on a course of action based on morals and autonomous moral
agents should not leave moral decisions to authorities or experts.
Autonomous agents are entities
in an environment that sense and act on that environment to pursue a
purpose. That purpose comes from its programmed drives and goals.
An example of non-human biological agents would, of course, be plants
and animals. Non-biological examples would be autonomous robots
and autonomous machinery.
2.2 Autonomous Robots
and Autonomous Machinery
Autonomous robots and machinery
are made to perform harmlessly in an environment without continuous
human guidance. This might include adapting a new preprogrammed
strategy to adjust to the surroundings. The human analogue begins
with a computer (brain) with a program and logic gates (neurons).
Input is received through the senses by seeing (laser, radar, etc.),
hearing (sound and voice recognition etc.), smelling (vapor sampling
for molecular profiling technologies), tasting (particle sampling technologies),
and feeling (thermal etc.). Those capabilities can be augmented
by television, radio, internet, and satellite positioning, etc.
Proprioceptive sensors analyze
one's own internal status and are necessary for maintenance.
Humans need proprioceptive sensors to balance, i.e. muscles, the vestibular
in the inner ear to track the position of the head, and the eyes.
Robots and machinery could proprioceptually determine if a battery source
is low and needs charging or whether it has parts that are overheating.
A simple decision tree can
be created from input as indicated in the following basic examples.
-or- scheduled duration
--Monitor circuits (sensors)
--OK -or- not OK
--If not OK repair -or-
--If repair self-solve
-or- not self-solve
--If not self-solve seek assistance
ON/AWAKE MODE--On from demand
-or- scheduled duration
--On function check circuits
--OK -or- not OK
--If not OK (same process as
--If OK proceed to task
-or- not proceed to task due to wait mode
--No problems in tasking
--If problems (same steps as
if not OK and complete task when the problem is solved)
--If no problems complete task
and continue ON MODE -or- start OFF MODE
--If ON MODE begin new task
-or- wait mode
--If wait mode refuel
-or- not refuel
George A Bekey, sometimes called
the father of autonomous robots, would include the concepts of motion
in the biological analogue of robots such as swimming, crawling, and
walking. Whatever the machinery and its closeness to the human
analogue, it is not self-aware nor has freewill. Thus, even types
of computers that may handle information fusion, which humans use for
morals, cannot understand morality and cannot now or ever be autonomous
moral agents in the strict sense of the meaning.
3. Political Implications
--The idea that there are human
heuristics, or an art to discovery, indicates intellect. This
process is greatly used as an analogue for human design for autonomous
agents like robots and machinery. Those manmade creations use
human sensory analogues and cannot use the moral analogues. They
can never procreate nor have a personality of love and justice, grace
and mercy, hope and charity, humor and irony, or worship and joy.
Political decisions should take into account those limitations as finite
resources are allocated.
personality analogues of humans in the area of study of depressive personality
analogues. Matter (atoms and molecules) has no ability or reason
to communicate personality characteristics, i.e. love and justice, with
other matter unless a designed program is superimposed onto neurons
(logic gates) creating self-awareness and thought. The Being of
intellect that created humans and other biological autonomous agents
would likely have created humans as an analogue of Himself.[Gen. 1:26-28]
Such a being would be a God of love.[1 John 4:8, 16] It is described
in the Bible that He commended His love toward us as sinners (noncompliant
with His Decalogue and its subsets) yet Christ (the only atonement and
perfect moral analogue) died for us.
[Rom. 5:8] It is only reasonable that man accept salvation and
love Him with all our hearts, souls, and minds and love our neighbor
as ourselves.[Matt. 22:37-40] Thereby a First World political
system can be made possible. Anything less would create inferior
and cruel Second World or destructive Third World political systems.
--The creator of humanity would
be the God of the first cause of all effects. He would have created
with purpose as man creates autonomous agents for a purpose(s).
Divine autonomous creations would not have unlimited freedoms nor should
finite man desire that destructive possibility. Man's limited
epistemological ability allows truth to calibrate the will and avoid
harm. Our wills could not be wholly satisfied and joyous apart
from knowing the Creator. Looking to other creations for eternal
hope would ultimately lead to disappointment but would point to their
creator--Almighty God. His Holy Spirit must open man's theological
eyes to see, love and bless or the logical concepts [2 Cor. 5:17] will
seem foolish, hateful and cursed.[1 Cor. 1:18] (All things being equal
i.e. IQ, health, education, chronological age, etc.) Following
God's purpose should be recommended because it leads to individual
and societal efficiency.
--Hope, from God, explains
why humans want the righteous underdog to win as told in the David (a
good boy) and Goliath (an evil giant) story.[1 Sam. 17] That intelligent
design is contrasted by the Darwinian evolution precept that there is
no hope or compassion and the evil giant, Goliath, should have prevailed
around 1015 B.C. over the good boy, David. Since we humans are
all, at times, righteous underdogs the negative evolution belief causes
cognitive dissonance for its followers due to the innate will for the
just to survive. Darwinian evolution also vainly opposes other
needs such as being loved by family and friends, and receiving affection
from acquaintances. Those needs must have positive reinforcement
for political cohesiveness and a stable economy. (The J Curve,
written by Ian Bremmer, poses a simple model for nation-state stability.
In an e-mail exchange between Bremmer and Dallas F. Bell Jr. during
August 2007 Bremmer said his model does not account for the variations
of gender, IQ, and theological beliefs to include eschatological beliefs.)
--Man's autonomous robots
and machinery cannot ever exceed their program potential. Neither
can man know or do more than their programs. Unlike the infinite
God's creations, finite mankind's creations are subject to the Law
of Unintended Consequences. Unintentional consequences always
result from a lack of infinite knowledge by the human creator/programmer
and may result from the creation not operating as thought when its input
was designed. All things consistent with natural law should, in
a sense, be considered as truth such as harmonic music and mathematics.
They were designed by God and are therefore moral. Then there
would not be any divinely created non-moral analogues. This concept
may help when defining international laws and pragmatic terms e.g. terrorism.
"Morally justified causes for violence (like freedom fighting) if
intentionally enacted upon defenseless civilians is terrorism and not
moral." (The preceding quote is an excerpt from an e-mail exchange
in August 2007 between a representative of the International Institute
for Counter-Terrorism and Dallas F. Bell Jr.)
Hopefully, the few implications
listed heretofore can lead to a greater self understanding and discovery
of man's intelligently designed environment.
RIGHTS RESERVED © 2007 DALLAS F. BELL, JR.----------