The
Q-AI Expert Technology System White Paper
This white paper describes and explains the Q-AI Expert technology system and its relationship to the expertise brokering technology architecture for networks, for which the Q-AI Expert architecture is the interface. The paper continues with a description of how to integrate these two architectures into a hybrid system that includes human intelligence, some applications for the hybrid system in various market areas, and it explains some key points about its deployment. The last section explains how to license the Q-AI Expert technology system and the technical support program that is available for it.
There are two architectures that make up the Q-AI Expert network technology system: One is for brokering human expertise over networks and the other is a human interface architecture.
The famous Turing Test created by Alan Turing (http://www.webopedia.com/TERM/T/Turing_test.html) says that the best way to know if you have a working Artificial Intelligence (AI) system is if a user cannot tell the difference between interacting with the AI system and interacting through a computer with another human being. Leaving aside all the various opinions for and against the validity and usefulness of this test, consider another interpretation: If a person believes they are interacting with a sophisticated AI system over a computer network, they may not be able to tell if a human expert is role playing with them as an AI persona. The reason for this is that over a network, it can be very difficult for a computer user to distinguish between the actions of computers at other nodes on the network and the human users of those computers. So unless video is used, it is hard for a network user to determine if the response they receive is from a computer at another network node or a human being who is impersonating an AI computer. (And even with video, in a few years computer animation will be good enough that it may be very hard to tell a real person from an animated talking head.)
This situation means that a Star Trek* like AI computer system could be simulated as a hybrid system that uses real, anonymous human experts to provide what state of the art AI systems cannot: A virtual consciousness that is capable of reasoning, common sense, and natural human language understanding. That is, real human consciousness would support the virtual consciousness of the hybrid system when whatever level of AI that is available in the hybrid system is unable to answer a user's question or provide the requested expertise. The human experts in the simulation system must remain anonymous, however, or the system would be like ordinary human to human communication over a network, such as email. It is precisely the anonymity of the human experts who role play the futuristic AI in the hybrid system that enables the users to think they actually are interacting with a powerful, and real, AI system. The degree of simulation of intelligence vs. the use of "just in time" real human intelligence will vary depending on the semantic information available on the Web. As computational epistemology produces more and more Web ontologies, less and less human intelligence will be needed to get the same result. Even in April of 2001, people were only beginning to realize the potential of Web semantics and ontologies (For details see Hendler - http://www.cs.umd.edu/~hendler/AgentWeb.html ). More recently, probably the best known system for representing Web ontologies is the OWL Web Ontology Language that was developed at Stanford by Deborah L. McGuinness and Frank van Harmelen of the University of Amsterdam . (See http://www.w3.org/TR/owl-features/ for details.)
Given this scenario, the items required to create a hybrid system like the one just described are at least, but not necessarily limited to, the following:
1) Two groups of users who are Seekers and Providers of information and expertise over a network: The Seekers and Providers may be human network users, and they may also be machine network users such as servers, PDAs, other AI systems, robots or automated vehicles, security systems, or any other automatic computerized systems that store human expertise and can function as Seekers and/or Providers.
2) A network connecting these users to a hybrid system consisting of ordinary computer applications, including ones that register information and expertise Seekers and Providers, automated pricing and negotiation applications (if the information and expertise provided by the hybrid system is not offered for free), security applications, any other applications that may be needed to make the hybrid system function legally, properly, and reliably, and the application that implements Q-AI Expert human interface architecture (or some other similar interface architecture).
3) Some form or level of AI applications that are part of the hybrid system such as a state of the art AI expert system, and an automated escalation system to elevate failed information and expertise requests from Seekers (requests that fail to provide the Seeker what is sought) to higher levels in the system where these failed requests are resolved.
4) A group of anonymous human experts who will be Providers of the information and expertise that the lower levels of the hybrid system are unable to provide.
5) A group of managers who monitor the hybrid system's operation to maintain its legality, security, and an optimal level of functionality for its users.
The interface for the hybrid system may be an implementation of the Q-AI Expert architecture or some other similar interface arrangement. The essential requirements are that whatever interface hardware and applications are used, they must preserve the belief by users that they are dealing with a futuristic AI system, and they must maintain the anonymity of the human experts who add the real, active intelligence to the hybrid system by role playing an AI persona as Seekers and Providers interact with each other.
To meet the first requirement of simulating a futuristic AI system, the Q-AI Expert architecture offers a range of personas such as the "cartoon" like "Quasi" example from our web site to more sophisticated personas, depending on the needs of the organization designing and building the hybrid system. The personas can be as simple as still graphic characters that communicate with users by means of text annotations and/or menus, to more animated graphics that use off the shelf animation and voice synthesizer applications, to the most advanced animated talking heads that may be available. The Information Technology (IT) team of the organization implementing the Q-AI Expert architecture to build their hybrid system can mix and match whatever existing hardware and applications best meet its needs and budget, or create their own interface personas and organize them into Augmented Reality or Instant Messaging systems.
To meet the second requirement of protecting the anonymity of the human experts who role play the AI in the hybrid system, the IT team of the organization implementing the Q-AI Expert architecture must take care to build a seamless software barrier between the hybrid system's users (Seekers) out on the network and the human experts (Providers), who may be either inside the hybrid system or out on the network. In addition to making the system easy to use, the main function of the interface technology is to mask the AI role playing human experts that provide the futuristic AI functionality, so as not to compromise the users' belief that they are, in fact, interacting with a powerful AI system. To meet this requirement, the IT team must adapt or modify whatever hardware and applications they use to work with the hybrid system's state of the art AI, escalation system, and other applications, so these applications provide appropriate and timely responses to users. Some of these responses may be to distinguish between human and non-human Seekers and Providers, to register Seekers and Providers in the system, to negotiate a price (if the information and expertise to be provided is not free), to refuse service if a Seeker or Provider asks for or offers something illegal, to communicate the information or expertise the Seeker has ask for when it is available, to ask the Seeker clarifying questions to narrow an information search or for escalation purposes, to tell the Seeker the search will take some time and they will be contacted later, and so on. These responses and questions can come from a database created for this purpose, or after escalation, from the anonymous human experts themselves.
Obviously, before any design work begins on the hybrid system, a detailed needs analysis should be performed to determine a snapshot of the current system in the domain of the project for the organization planning to implement the Q-AI Expert system, an identification made of the desired state of the system when the project is complete, a determination made of the changes required to migrate the system from the current state to the desired state, and a detailed design plan written that specifies both how the migration will be accomplished and how the final system design will be deployed and integrated into the wider domain of the organization as a whole.It should be a central theme of the design document that both the expertise brokering and the Q-AI Expert architectures must be seamlessly integrated into the overall hybrid system design model by the IT team building and implementing it. Therefore, it is highly recommended that the complete hybrid system should be designed and modeled using computer system design tools such as those for Computer Aided Software Engineering (CASE), in order to ensure the overall integrity and consistency of the mechanistic parts of the system. In addition, since human experts and managers are an integral part of the hybrid system, it is also highly recommended that some form of quality management system such as ISO 9001:2000 be employed to manage the actions of the humans who role play the AI personas to ensure and maintain "best practice" for the human behavioral aspects of the hybrid system. The following list contains some more specific design requirements for the hybrid system:
1) A quality management system like ISO 9001:2000 requires that your organization's core business processes and personnel information must be documented and stored in a business management system database and its index made available to the Q-AI Expertise brokering system. This task is usually not part of the IT team's job in building a new system, so be advised that it will be helpful to find a team member who has quality management expertise.
2) The roles of whatever number of human experts, processors, and managers required to operate the system must be defined, along with a work schedule to ensure these people are always available when the system is in operation.
3) The type and complexity level of personality simulation for the system's interface must be defined and created by the IT team from the range of options defined above and incorporated by them into the hybrid system's design. Therefore, a task analysis must be performed to identify which tasks will be automated, which tasks will be manual, the overall task process flow, what application or person will perform each process, and all control issues, such as how lower-level automatic tasks which failed to provide Seekers with requested information or expertise will be escalated to higher level state of the art AI applications or to the human experts and managers in the system.
4) A registration system for expertise Seekers and Providers must be created to build a database of the Provider's expertise and Seekers' needs. The real names and other information about the people who function as part of the hybrid system will need to be available for financial, legal, and security reasons, but it must be only available to key human managers to maintain the anonymity of the Seekers, Providers, and the human managers who control the system.
5) An automated matching system must be created to identify easy matches between Seekers and Providers. This application must be linked to or part of the escalation system that escalates failed requests to higher level AI applications and ultimately to human expertise Providers. A detailed set of escalation measurement criteria and procedures will also be required for this matching application. In addition, an automated screening system must be created to evaluate all automatic matches based on organizational policies, and a set of procedures and measurement criteria created in the quality management system to ensure that matches made by the humans in the hybrid system are also consistent with organizational policies.
6) An automated invoice and payment application must be created if the expertise brokering is not a free service.
7) During testing, the system must undergo a security and efficiency evaluations to get baseline measurements and define the state of best practice. On a regular basis after the system is up and running, further testing and improvements must be made to reach and maintain the state of best practice.
Once all of these requirements have been met in the overall design model, the code for the hybrid system can be written, tested, and the system as a whole should them be pilot tested. At this point, the Q-AI Expert system will be ready for deployment.
The procedures for deploying the Q-AI Expert Architecture in a given organization should be specified in the hybrid system design plan. Many of the procedures will obviously be specific to the organization doing the deployment. However, there will be a need for some general procedures in the deployment process.It is highly recommended that the source of those general procedures be the same quality management system used to ensure the best practice in the operation of the system as a whole, such as the ISO 9001:2000 standards.
In addition, a strong emphasis on training is suggested for everyone involved with the deployment and operation of the Q-AI Expert system. The specific performance based learning objectives needed to develop the training courses required can be determined from the system design plan and the course audiences, and the training courses themselves can be developed in parallel with the system as part of the overall project. Ensuring a high level of knowledge and skill on the part of all the key people who work with the system during and after its deployment will also go along way toward ensuring the system's ultimate success.
There are a number of market areas where brokering expertise can provide a very valuable resource. Some examples are:In Business
There are many areas in a large business where brokered expertise can be extremely profitable to an organization, expertise that is already available in the minds of an organization's people, but uncataloged and untapped. For example, many potential expertise Providers in a business organization, such as its executives, middle managers, engineers, scientists, staff employees, vendors, or even customers posses volumes of knowledge and expertise both in and outside their job roles, but that fact may not be known to the Seekers of the information or expertise because there is no index of it and means to access it. In addition, even if such access were available, it might not be possible for Seekers to ask for expert advice from many of these Providers because there are organizational political situations that would make it uncomfortable or inappropriate for an executive or customer, for example, to share information or expertise unless it could be done anonymously. Q-AI Expert technology can make it possible for organizations to broker their Provider's expertise anonymously and when appropriate, and it can enable a business to tap a new resource.
For Telephone Answering Systems
In recent years, it has become almost impossible to call an organization for help and talk to someone who can provide expert advice without fumbling one's way through endless menus of choices in automatic telephone answering systems, menus that almost never seem to fit the problem one has. Q-AI Expert can improve this frustrating situation by using simulated personas instead of databases of menus and by efficiently escalating callers to human experts role playing AI personas. The overall effect can be faster more friendly service to callers who are desperate to get some small piece of information or expert advice from an organization to solve a pressing problem they have.
By Providing Quick Access to Expertise Over Networks
There are many individuals who need information or expertise as evidenced by the popularity of search engines and other similar sites on the Internet. However, using Internet web sites is often difficult or confusing and sometimes produces frustratingly few good results. To be able to access what seems to be a Star Trek like computer that offered information and expertise for a reasonable cost would very likely be in high demand. Q-AI Expert technology can not only provide such a service, but at the same time, it can create an opportunity for millions of potential Providers of expertise and information to offer their services where they would have no effective way of doing so on their own.
Brokering Expertise for Technical Support
We live in a technological age, and there is no end to peoples' need for answers to their problems with technology. Since the Internet connects millions of people together, it offers a powerful medium for brokering the technical expertise of the Providers who have it to the Seekers who need it. Q-AI Expert technology can provide the means of bringing these two groups together in a practical and efficient manner.
For Cool New AI Network Games
A wide variety of games are now able to be played over the Internet. Q-AI Expert technology can enable a combination of ordinary network games, AI enhanced network games with various character personas, and seemingly very advanced, futuristic AI games enhanced with very human like character personas masking anonymous, real people who role play AI or robot characters. This arrangement can offer a whole new range of exciting possibilities in the game playing market for the Internet.
For Defense Applications
Modern technology is changing the battlefield quickly. The U.S. Army's Land Warrior system already links platoons of soldiers together with each other and their commanders within a network and connects their GPS systems with battle field maps. There are other functions that enable them to send pictures and other data to their command center and receive updated orders. However, no matter how good their training, unless they have seen battle many times, their experience and expertise can be limited. Q-AI Expert technology can make the battle tested expertise of more experienced soldiers available to freshly trained soldiers or those with less experience through the battlefield network. And, by adding expert human translators to the system (who can be located anywhere in the world and linked via satellite), Q-AI Expert technology can augment extant computer based translation systems by escalating complex or difficult exchanges to real people for translation. Of course, given the way Q-AI Expert technology works, to the soldiers and the person being interrogated, the system would seem like a universal translator.
For Augmented Reality (AR) and Instant Messaging Systems
Over the next few years as computer components become smaller and more powerful, AR systems will become more an more prevalent (See the April 2002 "Scientific American" http://www.sciam.com/techbiz/0402feiner.html ). In addition, many AR systems will include Instant Messaging as an integral feature, using AI personas as an interface. Both for defense and civilian uses, AR will offer faster and easier access to network resources than ever before. This means that users will exhaust the preprogrammed responses of network systems faster than ever and become frustrated even sooner than they are now, if the automated responses of network systems are unable to provide users what they seek. Q-AI Expert technology can make AR systems much more useful by providing nearly instantaneous access to anonymous human experts anywhere in the world, experts who can provide what the users need on a timely basis.
If you decide to license Q-AI Expert technology, the first step is to email us so we can negotiate a license agreement. Once the license agreement is approved, we will work with your IT team to support them as they develop the design for your system.In addition, we can design a custom training program in parallel with your development project, so that training courses are available at the time your system is deployed.
< Back to Q-AI Architecture page
< Back to DLF Simulation Technology page
© Copyright 1998 - 2007, Blue Oak Mountain Technologies, Inc. All Rights Reserved
* STAR TREK and related marks are trademarks of Paramount Pictures Corporation.
Welcome Page | About Us | Testimonials | FAQs