R. Falcone et C. Castelfranchi, The human in the loop of a delegated agent: The theory of adjustable social autonomy, IEEE SYST A, 31(5), 2001, pp. 406-418
Citations number
19
Categorie Soggetti
AI Robotics and Automatic Control
Journal title
IEEE TRANSACTIONS ON SYSTEMS MAN AND CYBERNETICS PART A-SYSTEMS AND HUMANS
In this paper, we refer to social autonomy in a collaborative relationship
among agents based on delegation and help. We address the problem of adjust
able autonomy, i.e., we discuss when modifying the assigned/received delega
tion entails a corresponding enlargement or restriction of autonomy and at
which level. We stress in particular the role played in autonomy by: 1) the
degree of "openness" of delegation (execution autonomy), 2) the allowed in
itiative in (re)starting negotiation (meta-autonomy), 3) the degree and kin
d of control (feedback+intervention), and 4) the strength of delegation wit
h respect to interaction. We show how the adjustability of delegation and a
utonomy is actually "bilateral," because not only the user (delegator, trus
tor, client) can adjust the autonomy of the agent (delegee, trustee, contra
ctor), but the agent can also have (cooperative) reasons for and the abilit
y to change the received delegation and modify its own autonomy in it. Adju
stment is also "bidirectional" (from more autonomy to less autonomy, or vic
e versa), and multidimensional. Finally, we analyze some reasons for modify
ing the assigned autonomy and show how the adjustment of autonomy depends-o
n the delegator's side-on a much less than crisis of trust much greater tha
n; vice versa, the delegee's adjustment of its own autonomy depends on some
disagreement about the trust received from the delegator, and, in particul
ar, either a higher or lower confidence in itself or in external circumstan
ces. Some preliminary hints about necessary protocols for adjusting the int
eraction with agents are provided. This work is neither directly prescripti
ve nor simply descriptive. It is aimed at providing a theoretical framework
, i.e., the conceptual instruments necessary for analyzing and understandin
g interaction with autonomous entities.