Artificial Intelligence: Benefits and Pitfalls
Dr. Shrikant V. Sonekar
Artificial Intelligence (AI) is succeeding at all the fronts of technology, right from speech interpretation and recognition interface to self-driving cars. It is one of the top-most technologies having huge potential in the present times as well as the future. In broad terms, Artificial Intelligence is the ability of a computer program to learn and think. The opportunity for innovation, modernisation and development in AI is enormous and it will keep changing the world in diverse ways in the future. Involvement of AI is everywhere, right from space exploration to advancements in defence systems and more. The technology is expanding progressively, and it has the potential to be more intelligent than ever. The phrase "human error" came into existence because humans make mistakes as and when they learn. On the contrary, computers, if properly programmed, don't make mistakes. When AI is used, decisions are taken from the previously gathered information applying a certain set of algorithms. So errors are reduced and accuracy is achieved with a greater degree of precision. AI is not restricted to only robots with human characteristics, it encompass anything from Google's search algorithms to IBM's Watson to autonomous weapons.
What is Artificial Intelligence?
In broad sense, Artificial Intelligence is a branch of computer science wherein smart machines are built in such a manner that they respond and behave like humans. The purpose of building an intelligent system, machine or robot is to support the decision making process which is completely analysed, based on the data that is available with the organization. The entire process of building an intelligent system is same as that of humans blending the information and providing suitable decision. The only difference is that in case of AI, the decision making is reached by analysing tons of generated and gathered data. Technocrats have categorized AI into two parts - Narrow AI and General AI. The former is designed to accomplish basic tasks or to outperform the basic human task such as facial recognition, playing chess, solving equations, internet searches or only driving car. The later one is meant to design cognitive, intellectual, perceptive and rational task.
Many vocations involve performing repetitive works like sending thank you email, verifying certain documents for errors etc. For example, verification of documents for processing loans, etc. The feature of AI Cognitive Automation helps to speed up the process of verifying the documents. Some of the highly advanced organizations use digital assistants to interact with users which save the need for human resources. Digital assistants are also used in many websites and mobile applications to attend to users' queries. We can chat with them about what we are looking for. Some chat bots are designed in such a way that it becomes hard to determine that we are chatting with a chat bot or a human being.
Artificial Intelligence incorporates both Machine Learning and Data Science
AI has many inferences, depending on the specific industry it is used in. Most often, when people say AI, what they generally mean is Machine Learning. The ability of a computer to imitate human behaviour is actually only the basic function of AI. Machine Learning and Deep Learning are the two subsets of AI. Machine Learning enables systems to make predictions or take decisions or learn from past data or experiences. It works on the algorithms which learn by its own using stored data. It allows computers to "witness" human behaviour through the intake of data and then undergo advanced processes to analyse that data and identify patterns within it. It uses a gigantic quantity of organised and partly organized data so that a tool or model of Machine Learning produces precise result or give predictions based on that data. Whereas, AI works on all category of data i.e. organized, partly organized and unorganized data. Machine Learning is the link between Data Science and Artificial Intelligence for the very reason that it is a process of learning from the data over time, while AI is the methodology that helps data science in accomplishing the goal of getting results and solution for a specific problem. Machine Learning is the technology which helps in achieving the set goal. Data science is not exactly a subset of Machine Learning but it uses Machine Learning tools to analyse data and make forecasts for the future. Data Science combines tools and methods of Machine Learning with big Data Analytics and Cloud Computing to enable the algorithms for predictions. Data Science is purely an applied application of Machine Learning with full attention and importance on solving real and material world problems
Benefits of Artificial Intelligence
i. Bright and versatile career opportunities: The amalgamation of Artificial Intelligence, Machine Learning and Deep Learning provides thoughtful path for individuals to make innovations. It also helps in terms of productivity and accuracy which ultimately helps in reshaping the world. We all have heard that AI is ready to replace a lot of jobs that humans do. Yes, it is true. Artificial Intelligence is one of the emerging technologies making its mark in every industry ranging from fashion to finance. It is creating more than 100 million roles in all the major sectors. AI has become intrinsic to key sectors like automobile industry, smart home devices, online customer support, security surveillance, retail, healthcare and many more. Therefore, AI provides bright and versatile career opportunities. Innovating and operating AI systems not only give you a decent salary but also a dynamic career trajectory. Job profiles can range from Machine Learning Engineer, Software Engineer, Hardware Engineer, Research Engineer, Business Intelligence Developer and even Data Scientist. Big employers like Google, Amazon, and Microsoft hire AI experts. People always think that AI draws its attention only in computer or space related industries whereas the reality is that Artificial intelligence has got everything in the bag to stand out from the crowd.
ii. Automation and smart decision making: AI technology has ensured a lot of progress in the field of automation and is cited as one of the most remarkable benefits of modern technology. It has had momentous impact in the field of communications, transportation, consumer products, and service industries. Automation not only concentrates on increasing productivity but also allows more efficient use of raw materials, improved product quality, reduced lead times, and superior safety. AI technology can coordinate data delivery, analyse trends, develop data consistency, provide forecasts, and quantify uncertainties to make the best decisions for the company.
iii. Aiding medical sciences and solving complex problems: AI has made enormous contribution in the field of healthcare. Remote patient monitoring technology, monitoring the progression of contagious diseases, prediction of future effects and outcomes, prompt suggestion for treatments, diagnoses of certain health condition without requiring the patient to visit the hospital are some of the key benefits of AI in medical fields.
iv. Research and data analysis: AI helps in analysing the data in more efficient manner. The technology helps to create predictive models and algorithms to process data and understand the impending outcomes of different trends and scenarios. The advanced computing features of AI help in speeding up the processing and analysis of data for research and development.
v. Managing repetitive tasks and ensuring continuity: AI powered robotic automation process helps in removing unnecessary repetition of tasks as well as performing repeated tasks. As risk management heavily relies on data management and analysis, AI-powered tools can help organizations to respond to the crisis proactively. AI technology not only helps companies to make critical decisions but also prepares them for any emergency to ensure business continuity.
Pitfalls of Artificial Intelligence
There are two sides to every coin. We, as humans, need to use our wisdom to utilize the positive side of any technology for creating a better world. There are many myths and theories that promote the fear of "rise of the robots." As on date, there is not a single application of AI which that can destroy or dominate humanity. Ultimately, it is a human driven resource made for the benefit of mankind. Nevertheless, the concern remain as to what will happen if the human quest for AI leads to an AI system so strong that it encompasses the human brains in all reasoning, intellectual and perceptive tasks?
In 1965, British Mathematician and Cryptologist I J Good, rightly pointed out that designing smarter AI systems in itself is a rational and thoughtful task. Discovering and revolutionizing the new technologies like Superintelligence help the society at large to exterminate war, disease, poverty and this is the reason why AI became the biggest and most important human innovation. Such a system could hypothetically experience recursive self-improvement, raising an intelligence explosion leaving human intellect far behind, thus triggering the concern whether such system could intentionally or unintentionally cause great harm to the human civilization. We can only hope that researchers will prepare a counter technology to inhibit possible negative consequences and ensure that humans continue to employ AI to their benefit while evading pitfalls.
Most of the researchers working in the AI domain think that any AI is unlikely to imitate complex human emotions like love or hate. Therefore, there is very less possibility for an AI system becoming deliberately compassionate or malicious. Scientists predict two negative scenarios that are most likely to happen if an AI system is not utilized properly.
i. Devastation: Autonomous weapons are artificial intelligence systems and their aim is to terminate objects they are programmed to kill. If it is given in the hands of wrong persons, these weapons could easily cause mass fatalities. To ensure greater efficacy, such weapons could be designed to be extremely difficult to simply turn them off, causing human intervention to fail. This risk is one that is present even in narrow AI, but grows as the level of AI increases
ii. High expectation, low results: AI are generally designed for beneficial tasks but it cannot be guaranteed that the system completely achieves its goals or does not work contrary to its original objective. This can happen if we fail to fully align the AI's goals with ours. For example, if you ask an intelligent car to take you to the railway station as fast as possible, it might get you there in no time but may cause destruction on the way.
(The author is Professor in CSE Department, J D College of Engineering and Management, Nagpur, e-mail: svsonekar@jdcoem.ac.in) Views expressed are personal
Image Courtesy : Google