The Social Impacts of Intelligent Agents on Internet Use

1
The Social Impacts of Intelligent Agents on
Internet Use

• Alexander Serenko Mihail Cocosila
• McMaster University, Canada
• AoIR, Toronto 2003

2
Agenda

• 1. The Purpose of the Study
• 2. What are Intelligent Agents?
• 3. Why Intelligent Agents?
• 4. Social Impacts of Intelligent Agents on
  Internet Use
• a) first-level impacts
• b) second-level impacts
• - user privacy
• - economic impacts
• - user confidence and trust
• - online habits and behavior
• - Web performance
• - online social interactions
• - online service providers
• 5. Conclusions and Directions for Future Research
• 6. QA

3
Introduction

• The purpose of the study
• To present a non-technical overview of
  intelligent agents technologies
• To identify and classify several possible domains
  of the social impacts of intelligent agents
  Internet use
• To briefly show the reasons why intelligent
  agents should ? (or should not ?) exist
• To suggest directions of future research
• It is the first attempt to discuss social
  implications of intelligent agent technologies on
  Internet use

4
What are Intelligent Agents?

• Metaphor of intelligent agents
• non-human electronic assistants
• a newer form of software
• possible to implement since the end of the past
  century
• new dimensions in the Internet era
• History of the concept
• introduced by John McCarthy in the 50s
• coined by Oliver Selfridge at the MIT Lincoln
  Laboratory
• agent visionaries in the 80s
• classic works in the 90s
• human-agent cooperation metaphor after 2000

5
What are Intelligent Agents?

• Intelligent agents as an ascription
• characterized in terms of what users ascribe,
  attribute, or assign agents to be
• agent exists only in the minds of people if
  individuals believe that they are delegating
  tasks to a particular software entity, this
  application is considered an agent
• This paper follows a description approach
• An intelligent agent is a software entity which
  is
• continuous (long-lived)
• autonomous (independent)
• reactive (adapts its behavior under the changes
  in the external environment), and
• collaborative (collaborates with other agents or
  electronic processes)

6
What are Intelligent Agents?

• General types of intelligent agents
• user agents
• assist human users by interacting with them
  directly, knowing their preferences and
  interests, and acting on their behalf
• Examples personal assistants, news editors,
  electronic shoppers, and Web guides
• service agents
• collaborate with different parts of a complicated
  computer system and perform more general tasks in
  the background being invisible for human users
• Examples Web indexing, information retrieval,
  and phone network load balancing agents

7
Why Intelligent Agents?

• Intelligent agents allow software users to
  utilize the indirect management approach rather
  than the previous less efficient direct
  manipulation method
• Direct manipulation approach people explicitly
  indicate all tasks the application should
  perform, monitor the process, observe results,
  and intervene when necessary
• Indirect management concept users are able to
  indirectly manage agents rather than directly
  manipulate objects agents, in turn, operate
  objects on users behalf and report back only
  final results thereby hiding tasks complexity
• Important while dealing with complex,
  heterogeneous systems, and unpredictable systems
  such as the Internet

8
Social Impacts of Intelligent Agents

• Every technological invention has dual effects
  on an individual, society, organization, group,
  or other social entity
• First level effects
• presumably positive outcomes
• always anticipated by technology inventors
• usually considered justifications of investments
• Examples dramatic cost reduction, improved
  quality, or the introduction of new products and
  services
• Second level effects
• can never be totally envisioned by researchers,
  innovators, and entrepreneurs
• natural consequence of altering peoples
  behavior, creating new values, expectations, and
  norms

9
Social Impacts of Intelligent Agents

• Web User Privacy
• Extremely important for proper functioning of
  modern society characterized by skyrocketing
  information exchange rates and availability of
  information collection tools and techniques
• Two basic ways intelligent agents obtain
  information about people
• Individuals may voluntarily express their private
  information when they first start using agents
• Intelligent agents work in the background by
  monitoring constantly all users activities such
  as surfing pattern, purchasing behavior, and
  search engine results

10
Social Impacts of Intelligent Agents

• Web User Privacy
• Your digital Mini Me
• Example Shopping bots (product assistants,
  shopping guides, intelligent portals, consumer
  oriented price comparison engines, auction
  watchers, book finders, intelligent online
  catalogues, and bargain finders)
• Improper use of people personal information may
  seriously undermine trust to intelligent agents,
  the Web, and online shopping

11
Social Impacts of Intelligent Agents

• Economic Impacts on Online Businesses
• Transform the existing online markets into the
  condition of perfect competition eventually
  predominantly price oriented
• The economic value of online information becomes
  one of the most valuable intangible assets
• Example Shopping bots (or shopbots)-price
  comparison shopping agents
• Increase the bargaining power of customers who
  may efficiently locate the best item price on the
  Web
• Transform the whole online shopping industry
  markets will move closer to the perfect
  competition model where the long-run cumulative
  profit of all sellers is a zero
• Drive sellers, in turn, to rely on intelligent
  agent real-time pricing technologies, or
  pricebots
• In conjunction with pricebots can theoretically
  engage virtual business in lose-lose price wars

12
Social Impacts of Intelligent Agents

• User Confidence and Trust
• The degree of user confidence in the quality of
  agents performance
• Does the agent present relevant information?
• Can it potentially find the best deal? and,
• Is the user able to exploit the full capabilities
  of the intelligent agent in order to obtain the
  desired results?

13
Social Impacts of Intelligent Agents

• User Confidence and Trust (continued)
• The level of user trust to an agent, which some
  people may associate with trust to many online
  services
• Does the agent search for the best deals on the
  whole Internet or only visits the sites of those
  online vendors who paid a subscription fee to a
  respective shopping bot designer?
• Can the agent avoid traps of the vendors trying
  to attract such type of intelligent agents?
• Would the intelligent agent be able to
  discriminate between serious online vendors and
  fake vendors or fake best offers?
• Would the intelligent agent be able to find the
  best deal from a more complex view rather than a
  pure price-only perspective?
• Website owners concern over various intelligent
  agent activities

14
Social Impacts of Intelligent Agents

• Online Habits and Behavior
• In an effort to save time, people will tend to
  reduce their efforts
• for information searching
• people will risk missing important information
  otherwise acquired by do-it yourself experience
• software entities can not always make the
  necessary associations to follow other branches
  of information retrieval because intelligent
  agents lack common sense
• Having intelligent agents perform an Internet
  search and selection on behalf of a human user
  may encourage substitutes of culture
• people will tend to have the intelligent agents
  do the entire job and receive ready to be
  digested and consumed information
• this may result in losing rather than acquiring
  knowledge if agents become too excessively
  exploited on the Web

15
Social Impacts of Intelligent Agents

• Peoples Web Performance
• The presence of intelligent agents in the form of
  personal assistants will make websites more
  accessible and even friendlier for those people
  who may be very good specialists but, on the
  other hand, have lower computer and Internet
  skills
• Intelligent agents may expand the usage of
  various capabilities offered by many Web sites
• Example Intelligent agents would be able to
  educate users on portal resources usage and offer
  real-time navigational help

16
Social Impacts of Intelligent Agents

• Online Social Interaction
• Modern work environment is becoming associated
  with the idea of knowledge portals embedding
  intelligent agents
• Present tendency of creating virtual and
  networking organizations where individuals work
  remotely from home and communicate with each
  other electronically
• Immediate consequences extra costs savings
  (commuting expenses and office space), more
  convenient work pattern
• Unpredictable consequence people may replace
  direct social interaction with mediated
  interaction which contradicts social work habits
  people have been developing for millennia
• Massive and indiscriminate use of software agents
  to facilitate work processes through the Internet
  may have significant social drawbacks in the long
  run

17
Social Impacts of Intelligent Agents

• Online Service Providers Activities
• The use of intelligent agents will change some
  peoples job tasks and activities, which will
  affect their job and life patterns
• People will have to comply with new job
  requirements
• Example Intelligent agents, in context of a
  library portal would perform most of the
  repetitive and tedious tasks of librarians
  librarians will be able to do more qualitative
  work requiring human innovative thinking
• Intelligent agents may potentially restructure
  the current online IT labor market and facilitate
  the creation of new jobs
• Example The boom in agent technologies will
  increase demand on information technology (IT)
  personnel, agent researchers, and agent oriented
  programmers (AOP)

18
Conclusions and Directions for Future Research

• The authors view
• Two extremist views on the usefulness of
  agent-based computing
• intelligent agents should not be included into
  Web-enabled applications
• agent potential should be fully realized, and
  people should be totally eliminated and replaced
  by machines and software
• The authors, however, suggest the third approach
  where intelligent agents should be included into
  online systems only under appropriate business
  and technical conditions and after thorough
  considerations of both the projected benefits and
  the unexpected social impacts of agent on
  Internet usage

19
Conclusions and Directions for Future Research

• Future research
• present investigation touches upon only a few
  unforeseen social effects
• each identified social effect provides a fruitful
  research field which may be further explored in
  the form of empirical investigations and
  conceptual discussions
• the existing social theories may be used to build
  a framework of social impacts of agents on Web
  usage

20
Conclusions and Directions for Future Research

• Final thoughts
• The intrusion of innovation into all aspects of
  peoples lives is especially difficult to predict
  in the early stages of technology development
• It is impossible to say whether intelligent
  agents will be able to fully perform the tasks
  that visionary researchers have predicted
• Future researchers should be ready to recognize
  those unanticipated effects

21
Questions?

• Contact
• Alexander Serenko
• Ph.D. candidate (MS/S) at the Michael G. DeGroote
  School of Business, McMaster University,
  Hamilton, CANADA
• serenkav_at_mcmaster.ca
• Mihail Cocosila
• Ph.D. candidate (MS/S) at the Michael G. DeGroote
  School of Business, McMaster University,
  Hamilton, CANADA
• cocosim_at_mcmaster.ca