The term Artificial Intelligence (AI) refers to a technology whereby intelligent agents are studied and created (Poole and Mackworth 9). In this case an intelligent agent is a structure that senses its surroundings and acts in such a manner as to maximise its chances of success. The origin of AI was in 1956 where John McCarthy described it as a science of creating thinking machines (intelligent machines). The technology comprises of eight different areas that are currently researched on and are wide in their own rights. The fields are: expert systems; decision support and planning; speech recognition; natural language processing; computer aided instruction; computer vision; and robotics (Poole and Mackworth 11).
The technology was founded in 1956 at Dartmouth College by John McCarthy and others who together wrote programs that were amazing at the time. The algorithms enabled computers to solve mathematical problems, provided logical theorems, and speak English. This sparked an interest in the field whereby by 1960 the Department of Defence was funding research in this area in addition to laboratories being set up globally. The advancement continued up to the 21st century where the technology is used in logistics, medicine, data mining and other industries. According to Poole and Mackworth, “the great advancement in AI research was as a result of increasing computation power of computers, linkage of AI with other fields, and commitment of researchers in discovering new trends in the field” (17).
Creating intelligence in machines is considered a problem and the problem is broken down in units called sub-fields. The sub-fields consist of traits that researchers would like the AI system to display. The traits that received the most attention are the ability: to deduct, reason, and solve problems; represent knowledge; have motion and manipulation; social intelligence; creativity; perception; and natural language processing. However, Poole and Mackworth asserts the general limitations that intelligent agents have through a term they referred to as “artificial stupidity” in the systems (22). The limitations he referred are: absence of common sense reasoning, ability to learn and explain, acquisition and maintenance of knowledge, and ability to support distributed expert systems.
Evaluation of AI is done using a test called Turing test that was prescribed by Alan Turing in 1950. The test allows all limitations of an intelligence agent to be tested, which at the current moment all agents fail. However, all is not over as they can still be evaluated on specific problems they do such as game playing and even solving mathematical problems. These tests are considered as ‘subject matter expert Turing test’ and provide attainable goals with an increased chance of positive results (Poole and Mackworth 26). The outcomes of AI tests are classified under four categories that are: optimal meaning it cannot be done better, example draughts performance; strong-super human meaning the performance is better than all humans such as chess performance; super-human meaning better performance than most humans; and sub-human meaning performance is less than that for humans.
This paper will critically examine AI as an emerging technology that has the potential of transforming industries. Analysis of trends and applications would be done in order to understand the importance of the technology. Moreover, two industries—automotive and telecommunication—will be analysed describing the impact of AI on them. This analysis will bring out the values of the technology and management approach that can be used in implementing it in industry.
Trends and Application of AI
Since the introduction of AI technology it has undergone three phases of development, which are: romantic, Ice age, and application period. The romantic period was experienced in the period 1956-1965 and was characterized by a product called the General Problem Solver that completed various mathematical tasks. The next phase was the ice age period 1965-1980 characterized by development of programming languages such as PROLOG and LISP, and discovery of logical models and mathematical languages using the Robinson solution. The last phase the technology has undergone is the application period experienced from 1980 to 2010. The phase was distinguished by usage of the technology in the military, industry, medicine and other services.
Currently, in 2011 the phase that AI is in is referred to as the Age of merging where research and application of the technology is done to come up with new systems (Polit 32). The leading countries in the high technology R&D are USA, Germany, Japan, France, Ireland, Finland, and China. The technology is not new to both firms and the world and can be shown where in Brussels and Washington all laws that are enacted depend heavily on expert based systems which are a segment of AI. The technology has provided the possibility of allowing 3-D voice commands for internet, radio and television, medical care and even telephones. Nowadays the educational systems are dominated by intelligent computers and telecommunication networks that aid learners more than they used using the old systems. The mergence and development of AI technology has resulted in development of other disciplines that are supported by the technology such as nanotechnology, biotechnology, and information technology.
AI techniques that have passed the research phase and are into the mainstream are too many to list. In the mainstream they are no longer referred to as AI but rather a phenomenon called Artificial Intelligence effect. AI effect is noticed in industries such as finance where AI systems are used to handle properties, stock investment; and organize operations. In the medicine industry the technology is used to provide medical information, make a staff routine schedule, and organized bed space. Furthermore, artificial neural networks in medicine are used for making decisions when coming up with diagnoses. In heavy industry, robots using AI have come in handy, often given jobs considered dangerous to man. In this industry Japan is a leader producing over 1.7 million robots globally according to the 1999 AI business survey (Polit 13). Other applications for this technology are in game playing, expert systems, natural language recognition, speech and language recognition, and computer vision and virtual reality.
In studying the increasing application of the technology to nearly all sectors of the industry it is becoming clear that the technology will become globally adapted in the future. Polit, forecasted that there would be an age of self reliance amongst artificial agents starting the year 2020 (56). During this age the intelligent agents would be able to handle their own scientific research, repairs and production. There would be a direct communication between humans and computers with certain legislatures recognizing human rights for some computers. In addition, the technology would ultimately lead into creation of new world that would be inhabitable by humans. This is through usage of intelligent networks that would propel moons—Phobos and Deimos—into the Martian polar caps making them human conducive.
Impact of the AI in Industry
From the previous section we have learned that AI technology has numerous applications in various sectors of the industry. With increasing developments and research in the individual disciplines, the same is happening to AI whereby a point is reached when there must be an association between the two. The impact of the technology in industry can be illustrated through analysing it with reference to the automotive and telecommunication industries.
I. Automotive Industry
The introduction of moving line assembly in the automotive industry enabled the industry to become a major incubator and innovator of Artificial Intelligence. The rapid emergence of numerous sophisticated vehicles is as a consequence of the growing complexity in the production and manufacturing, logistics, and supply chain departments all of which are considered the core of a modern automotive industry. AI technology is found in practically all areas in the industry beginning with vehicle on board systems through to value chain, designing, manufacturing, and after-market services (Polit 134). The focus of AI in this industry is to find more effective methods than normal engineering techniques.
A case study is the implementation of AI technology in Ford in order to cut costs and improve efficiency in the vehicle planning and production line. The GSPAS (Global Study Process Allocation System) was set up so as provide a standard methodology and business practices to be used in designing and manufacturing automobiles in the company’s plants globally. The system contains an embedded AI component called DLMS (Direct Labour Management System) that has two functions that are: improving process planning through achieving standardization at vehicle description process and providing a tool that applies standardized labour times which is essential during assembly; and provision of a framework that allocates operators the required work.
Implementing AI in one domain of Ford’s business has resulted in provision of leverage across other functional areas of the company. For instance, creation of successful applications in the manufacturing domain has resulted in creation of similar systems in the production, supply chain, and logistic domains of the company. Furthermore, the use of AI in Ford has proved that the technology can be incorporated into the business model and ultimately result in improvement of both quality and customer satisfaction.
II. Telecommunication Industry
The industry is significant with specific high technology applications that are specified with integration of business process and applications in IT. IT is used for promoting the competitiveness of the sector (Polit 178). Telecommunication environment is characterized by its ability for continued expansion in size, distribution, and specific fault-tolerant attributes. AI has been used in telecommunications since 1998, signifying more than a decade of influence of AI in the industry. Nowadays, expert systems are used with the aim of diagnosing complex equipments in an off-line mode (Polit 183).
In addition the need for telephony conversational applications offered by AI showed business value, and attracted the attention of numerous players in the telecommunication industry. In the early ninety’s, professionals forecasted that the speech market (including hardware, software, and services) would become multibillion industry in the future. As a result in the early 2000, there was an increasing trend towards commercial deployment of sophistication and push for new technology designed at large research labs like IBM and AT&T labs. This forced the industry to move from direct dialogue paradigms to the more sophisticated interactions.
A case study is where IBM successfully managed to launch the first commercial mixed-initiative solution with T. Rowe Price, a major mutual funds company, using AI’s natural language understanding and dialogue management technologies developed under the DARPA communicator program (Polit 187). This solution had an impact of being able to handle natural language queries, such as I would like to transfer my cash from bank A to bank B as well as resolving elliptical references, and allowing users to change the focus of the dialogue at any point in the communication (Polit 188).
In internet, semantic technology helps computers understand data through integrating large scale data sets. In addition, there is enablement of computers to deduce relationships amongst undefined data elements, hence ease in using searching applications. Semantic internet, that provides an avenue where machines can have conversations, makes the internet more intelligent. In being intelligent computers can analyze interactions between humans and computers, all data, and contents and links online (Polit 212).
III. Implementing AI Technology in Industry
Automotive and Telecommunication industry are all sub-fields of engineering industries, therefore in estimating the cost of introducing and implementing the technology in an engineering industry, there must be factors considered by the cost engineer. In this case one industry will represent the other as they are somewhat similar. In determining the cost of implementing AI technology in an automotive industry, Polit listed the factors as efficiency required, business strategy, market demand, type of AI technology, and industry standards (245). Efficiency in this case is a determinant where the efficiency level of production, manufacturing, and service required will ultimately determine the cost of implementing the technology. This is the case where if efficiency is required in all departments, then extra cost will be incurred than when efficiency is required in all departments in an automotive plant.
The business strategy of an automotive company will also determine the cost of implementing the technology. For instance ford had a strategy of expanding its operations globally and having more sophisticated brands. Therefore the strategy of the company will determine how much investment to be put on the technology. Market demand of certain types of goods and services will also influence the cost. In the automobile industry, there is an increased market demand of sophisticated vehicles which require AI technology in order to meet the demand. This demand will increase the cost in implementation of the technology. The last factor, type of AI technology is a major one where various players in the industry require certain specific types of AI technology. For example, a car manufacturer will require an AI technology that will assist in efficiency in the supply chain while another one in the production. Therefore in estimating costs you might find that one manufacturer will incur less or more costs than the other one.
Once the factors determining costs are understood and considered, then the next step is determining the costs that should be considered in implementing AI technology in an automobile industry. The costs to be considered are (Polit 246): research and design costs; testing costs; equipment costs, maintenance cost (life cycle cost); and operation cost. An automotive firm must have a research and development department where the technology can be researched on, designed and development. This process is costly and requires skilled personnel. After identifying the AI technology that is required then, the firm is considered to be ready for an artificial intelligence effect (AI application).
The management of a firm will have to consider the cost of testing the technology to see its effectiveness. Once that is considered then there is the cost of equipment and installation, maintenance and operation costs. In operation costs, the cost of labour is incorporated as running the technology requires highly skilled and experienced personnel. In conclusion, expert systems are used in the engineering industries where software that incorporate human intelligence in problem solving, or clarify uncertain issues that would require more than one human to be consulted (Poole and Mackworth 62). Therefore, costs in setting up the technology in the industry should be centred on incorporating an expert based system.
IV. Value of AI technology to Industry
The use of AI technology in the automotive industry has resulted in the reduction of cost in doing business. Even though, the initial cost is very high the long term benefits are worth the usage of the technology in automotive industry. With the introduction of moving line assembly in the automobile industry, the reduced cost of hiring labour and reduced accidental risks add to the values of intelligent systems.
Furthermore, profits are realized when there is increased production as a result of incorporated intelligent systems that are efficient in the assembly line which can even work for twenty-four hours.
Apart from usage of the technology in the assembly line of the automotive industry, it can also be used in production control and component manufacturing where quality control systems using the various algorithms enable firms prevent wastage of resources, produce quality products preventing recalls hence saving of costs in the long run. Furthermore, the system has come in handy in warranty management.
According to AMR research, the auto industry in USA pays $8.5 billion dollars per year in warranty claims which is approximately 1-3% of automotive profits (Poole and Mackworth 54). Therefore, the WAPS system—an AI technology—is essential as it uses the historical data of warranty costs and contained data using a measure called ‘warranty and goodwill’ rate as a foundation for predictive modelling. The predicted costs could be used in preparation for future costs hence help in planning.
The usage of expert systems in the automotive industry has also incorporated EIA (Environmental Impact Assessment) so as to comply with the current trend of green technology. This has impacted the environment positively with advantages being that the expert systems: help the user cope with large volumes of environmental management work; deliver expertise on environmental management to the non experts; enhance accountability of environmental management systems; and provision of a structured approach to EIA (Polit 142). Furthermore, with the use of intelligent agents there are reduced wastes that were associated with humans such as excretion and emissions that act as pollutants. They also tend to use less water and do not consume food all of which preserve the environment. The only negative aspect on the environment is the possibility of AI components being non-recyclable which again can be converted into recyclable parts.
With the introduction of AI in the automobile industry, there will be a considerable change in the way business is carried out. Most automobile firms will have robots in their assembly line, with all departments being efficiently run by intelligent systems that have replaced human functions. Furthermore, companies having that technology will have an advantage of maximum productions hence higher profits.
However, there needs to be legislations set up so as to: guarantee human jobs do not disappear; limit the number of AI technologies a particular firm has; and limit the maximum production per specific period in a firm. Automotive firms using AI technology in their operations have an opportunity of increasing their market share in the industry through the increased production of sophisticated automobiles. There is also an opportunity of being competitive in offering quality products that are defect free.
AI Risks and Disadvantages
AI technology is seen to destroy customary human jobs resulting to new forms of unemployment which is on the rise (Poole and Mackworth 22). Therefore because of less people working there is a possibility of the money supply perpetually reducing. There is also a risk of firms exploiting consumers of the basic services through job cuts, fines for delays in paying for their services, and refusal to supply adequate services and products in remote areas that do not yield profits as witnessed by the telecommunication giant Telstar (Polit 13). Risks due to lack of creativity may also arise where it is noticed that intelligent agents do not reason creatively but as they are programmed. Due to lack of creativity, problems would occur when disaster occurs in the firm which require human reasoning.
In addition to the risks, the disadvantages of the technology are that (Polit 14): it lacks human qualities such as reasoning; they can malfunction and do the opposite of what they are programmed to do; can corrupt the young generation; there is no filtering of information; and can be misused to cause mass destruction.
AI technology is an emerging technology and has the potential of transforming industries but only if it is implemented in the right way. The required management approach in successfully implementing this technology can be done through the following steps: 1. using a cost engineer to evaluate the cost of developmental equipment; 2. using a real customer problem; 3. Building a simple system or prototype; 4. Evolving the idea through frequent inputs from the customer; 5. Using small steps in implementing the system; 6. Identifying the value gained by the technology; and 7. improving the system continually.
Polit, Monique. Artificial Intelligence research and development. Amsterdam: IOS Press, 2006. Print. : 9+
Poole, David L., and Alan K. Mackworth. Artificial Intelligence: the foundation of computational agents. New York: Cambridge University Press, 2010. Print. : 3-65