Frequently Asked Questions
1-
What is an animated agent anyway?
An animated agent is a character (cartoon, human-like,
shape, etc.) that uses text, voice, and gestures to "communicate"
with the human user, thus serving as a visual interface between
the human being and the "cold technology". The use of animated
agents tries to make communication between the user and the computer
program more "human like", by providing an animated character
that seems to behave and act independently from the program. That
animated character (animated agent) offers the user help, guidance,
or support in the use of that program, as a human tutor or assistant
would. The agent's goal is to offer a friendly, helpful, human-like
interface that simulates real, live help, which would promote
a positive attitude towards that computer program and a higher
degree of self-confidence on the part of the user.
2-
Anthro-what? What on earth is that?
Anthropomorphism.
Animated agents come in different shapes and sizes, and they have
different roles, functions, and abilities. Anthropomorphism is
the degree to which the animated agent tries to resemble a human
being. For example, some programs geared towards younger kids
may use agents that consist of shapes or objects (a talking M&M,
for instance), others may have cartoon characters, and yet others,
geared towards an adult audience, may have animated agents with
a human-like appearance. In these last cases, those animated agents
are said to be "anthropomorphized" due to their high degree of
similarity with a human look-and-feel.
3-
How intelligent are the agents?
An
agent is as intelligent as a computer. This means, without some
heavy programming behind them, neither agents nor computers are
able to perform anything on their own. They depend on solid, sound,
bug-free programming to perform different tasks. So, a great-looking,
appealing agent, may be completely useless, because the programming
behind it is flawed and faulty. A poor-looking agent, on the other
hand, may have more capabilities, since it is the visible interface
of a strong program. As far as artificial intelligence goes, the
agent's intelligence will be limited to what research in computer
artificial intelligence affords. As said in the previous paragraph,
the agent is not intelligent in and of itself. It will only be
as "intelligent" as the technology used to support it. So, all
the limitations that artificial intelligence currently has to
allow interaction between human beings and computers, also apply
to interactions between human beings and animated agents (assuming
a specific animated agent is supported by artificial intelligence
technology).
4-
Can I carry a conversation with an agent?
Due
to the constrains that current technology has, a real-time, natural
conversation with an animated agent is limited in several ways.
Some agents have a pre-programmed set of context-specific expected
answers, which the user will receive as a response. Other agents
may rely on "artificial intelligence" technology, in order to
make "intelligent" choices as to how to continue and keep a conversation.
In any of these cases, though, current technology does not allow
for a real, natural conversation with an agent, the same way that
kind of conversation is still not completely possible with a computer.
Remember that the agent will only be as good as the computer technology
supporting it. A further constraint is whether communication is
written or oral. For written communication, a complex set of algorithms
can be used so that, given a comprehensive grammar structure is
provided, spaces can be filled in with the proper function words
(articles, prepositions, etc.) and meaning words (nouns, verbs,
etc.), so that acceptable sentences can be formed "on the spot"
for the agent to carry out a conversation. However, if communication
is oral, then a new set of elements, called "suprasegmental features"
(stress, rhythm, intonation) need to be considered. These features
affect entire phrases, clauses, and sentences, which make it almost
impossible to have pre-recorded individual words, and put them
together on the spot, expecting to have a fluent, natural phrase,
with natural inflection and intonation. As mentioned before, artificial
intelligence technology still has its limitations, so both written
and oral interaction will eventually reach a point where no flexible,
natural exchanges will occur, and the system will uncover its
weaknesses to engage in natural conversation.
5-
So, what's the usefulness of the agent?
Even
with their limitations, animated agents may perform a number of
roles (desktop assistants, personal assistants, presenters of
information, mentors, learning companions, pedagogical agents,
virtual reality environment assistants, etc.), where they would
perform a set of limited, well-defined tasks, with a good degree
of accuracy. In those roles, the main goal of the agent is to
provide a friendly, human-like interface to "hand-hold" users
through tasks and processes. As users have different degrees of
experience, expertise, and self-confidence, not all users will
see the agent with good eyes. An agent can be disturbing, distracting,
and annoying, for someone who already knows his or her way around
that specific computer program that features an animated agent.
But for other users, be it because they are not computer-savvy,
because they are still afraid of technology, or because they see
the computer as an "unforgiving, unhelpful, cold piece of technology",
the animated agent may be the element that allows them to have
a more "natural, human-like" interaction with that computer, so
that their level of self-confidence, and their feelings of being
supported and backed-up, are significantly increased.
