Trust, Control and Autonomy: A Dialectic Relationship - Trust Theory: A Socio-Cognitive and Computational Model - page 199

(renounce to surveillance), and when through surveillance? A detailed model of how and why
trust creates/increases trust is necessary in order to answer this question.
Should we make our autonomous agents (or our cyber-partners) more reliable and trustwor-
thy through responsabilization or through surveillance?
We will not have this doubt with artificial agents, since their 'psychology' will be very simple
and their effects will not be very dynamic. At least for the moment with artificial agents control
will complement insufficient trust and perhaps (known control) will increase commitment.
However, those subtle intertaction problems will certainly be relevant for computer-mediated
human interaction and collaboration.

7.1.10 Control Kills Trust
Control can be bad and self-defeating, in several ways.

There might be misunderstandings, mistakes, and incompetence and wrong interven-
tion by the controller ('who does control controllers?') (in this case Pr ∗ (achieve(g))

<

Pr(achieve(g) ).
Control might have the opposite effect than function ( kindA ), i.e. instead of improving
performance, it might make performance worse. For example by producing anxiety in the
trustee or by making him wast time and concentration for preparing or sending feedbacks
(case in which Trustworthiness ∗ (Y τ < τ

) ).
It can produce a breakdown of willingness. Instead of reinforcing commitment and willing-
ness, control can disturb it because of a bad reaction or rebellion, or because of delegation
conflicts (Castelfranchi and Falcone, 1998) and need for autonomy; or because of the fact
that distrust creates distrust (also in this case Trustworthiness ∗ (Y

) Trustworthiness(Y

τ ) < Trustworthiness

) ).
It can 'signal' (Section 6.3) to Y a lack of confidence and thus impact on (decrease) Y' s
self-confidence and self-esteem, negatively affecting his performance or commitment.

(Y τ

Here we mainly care about the bad effect of control on trust, which lets us see these dynamics.
As trust virtuously creates trust, analogously the trust of Y in X , that can be very relevant for
his motivation (for example in the case of exchange and collaboration), can decrease because
X exibits not so much trust in Y (by controlling Y ).

X is too diffident, does this mean that X is malicious and Machiavellan? Since X suspects so
much about the others would she herself be ready to deceive? Thus if X distrusts Y , Y can
become diffident about X .
Otherwise: X is too rigid, not the ideal person to work with. 15
Finally, if the agents rely on control, authority, norms they relax the moral, personal, or
affective bonds, i.e. one of the strongest basis for interpersonal trust. Increasing control pro-
cedures in organizations and community can destroy trust among the agents, and then make
cooperation, market, organization very bad or impossible, since a share of risk acceptance
and of trust is unavoidable and vital.

15 Control could also increase Y 's trust in X , as a careful person, or a good master and boss, etc.