Cloud Hosting

As the use of personal computers (PCs) became more and more widespread and now the proliferation of cloud and smart devices, numerous battles over turf broke out. These involved such issues as:

1. Which part of the organization controls the selection and acquisition of these devices?2. What procedures must be followed to control access to and the modification of corporate data bases?3. How should these devices and their software be networked together?4. Who is responsible for developing or acquiring new software?

Data processing and management information system (MIS) groups have found it necessary to modify some of their established procedures to deal with the challenges of PC technology. The intent of this modification is to support distributed processing on a network of small computers while retaining the overall responsibility for ensuring that the organization's corporate resources are used most effectively. As Al technology is more widely used, what will be the change in the role of these data processing and MIS groups? Will AI become just another part of data processing?

Numerous trade-offs are possible for assigning responsibilities for developing or utilizing Al systems. Should the existing MIS group supervise the development of information systems, or should a new in-house Al group take over that responsibility? Factors to be considered include:

1. The level of interaction needed between these systems and existing corporate data bases2. Familiarity with the organization's needs, procedures and existing data-processing systems3. Cost of equipping, training, and motivating a specialized Al staff4. Built-in NIH biases ("That's not our idea, just do it the same way we always have.")5. Attitudes towards working closely with "nonprofessional" or "hands-on" experts such as those on the factory floor or in customer service6. Requirement for new specialties7. Distinctions between development of systems intended to improve internal operations and development of new products or services8. The amount of EDP resources required to develop or run an AI application program

The IT groups certainly have had extensive experience in interfacing with many elements in the organization. However, they have not always been successful in completely understanding the needs of users or the methods used in accomplishing specific tasks. Although they may be familiar with computer technology, some MIS personnel are not suited for the level of innovative development required with the current state of artificial intelligence art. Conversely, they may have become by reason of previous experience much more realistic about scheduling and cost requirements. Finally, motivations and priorities may favour the establishment of a specialized AI group.

One person spent several hours with the members of a consulting group that specialized in the design of large database systems. The purpose of the meeting was to explore the commonalities and differences between AI and "conventional" data-base system practice. There were two interesting conclusions from the meeting: First, that the Al community was just beginning to learn what the data-processing community had learned long ago, and second, that the major difference was one of focus. The designer of a data-base system must ruthlessly focus on commonality, suppressing any individual differences. The designer of an Al system, on the other hand, gives the greatest emphasis on the individual and his or her needs.

As distributed computing power becomes more ubiquitous, it may be possible to embed individual support systems within the common whole. But there is also an opportunity for building distributed support systems that span the globe much more easily and can concentrate its support to areas of need when and where the need occurs

Excerpt from:
The Role of Data Processing Organizations in Artificial Intelligence - Toolbox