Information Technology Reference
In-Depth Information
Evidences for Trust
(55%)
Evidences for Distrust
(45%)
Evidences for Trust
(55%)
???????
(45%)
Figure 4.4
Evidence-based Trust matched with. a) Evidence-based Distrust and b) lack of knowledge
for trusting, but being inferior to another Agent (
Z
) is not winning.
X
trusts
Y
, she finds
Y
even
reliable and might entrust him, but - since there is also
Z
, who is even better - she will chose
Z
and (en)trust him. This is an 'insufficient' evaluation of
Y
but not really a 'lack of trust' in
Y
. Or better, it is a 'relative lack of trust' but not an 'absolute lack of trust'.
4.8 Trust on Credit: The Game of Ignorance
Trust (usually) goes
beyond
the evidences and certainty of well-grounded beliefs; in this it
consists of 'giving credit' (or not), that is, to believe even beyond the evidence.
Let us adopt a model
`ala
Dempster and Shafer,
8
where, between the evidence in favor
of a given eventuality that
P
(thus of the well established probability of
P
) (say 55%) and
the evidence in favor of
Not-P
(thus its estimated probability) (say 15%), there is a gap of
'ignorance' (say 30%), a lack of proof in favor or against
P
. First of all, let us observe that
with a given level of 'evidence-based' trust (say 55%) it is very different whether the rest, the
complement, is supported distrust or just lack of knowledge, just possibility, not probability
of
Not-P
(see Figure 4.4).
These two scenarios are psychologically very different: they induce opposite behaviors. For
example, in the second scenario the subject might suspend her judgment and decision while
waiting or searching for new data; she might perceive a strong uncertainty and be anxious.
While in the first situation she might feel able to take a decision (positive or negative).
This is additional confirmation of the fact that it is not sufficient to have just an index, a
number (say 55%, or 0.5) to represent trust.
It is important to stress that
with just one and the same level/degree of trust there may be
completely different feelings, decisions and behaviors
. Moreover, an 'insufficient' trust can be
due to quite different cases: factual distrust, negative evidence, or too much uncertainty and
ignorance.
4.8.1 Control and Uncertainty
As we just said, lack-of-trust because of insufficient information is different from lack-of-
trust due to supported low evaluations. In the former case, I can wait or search for additional
information. Sometimes, it is possible to adopt such a strategy as 'run time', 'work in progress',
and then start to delegate (trust is enough!) with the reserve of
monitoring
.
8
See for example: http://en.wikipedia.org/wiki/Dempster-Shafer theory
Search WWH ::
Custom Search