INTRODUCTION
A cognitive process displayed by non-human organisms, such as computer programs that carry out tasks resembling human cognition, is known as artificial intelligence (AI). AI is widely used in both business and consumer settings today because of its many advantages, including the ability to perform tasks as accurately as a human worker but much more quickly. Neural networks, deep learning, machine learning, computer vision, and natural language processing are just a few of the subcategories of AI techniques and software that are primarily utilised in Internet applications and smartphones.
The two main goals of AI and machine learning are to solve problems that people are unable to solve and to handle complicated issues that call for constant manual labour. Large data sources that expand over time can be effectively understood and insights extracted by machines, enabling firms to innovate and engage in revenue-generating activities. AI and machine learning may utilise data insights to forecast future behaviour, allowing companies to monitor important consumer behaviour and respond quickly to capitalise on occurrences the technology detects. Three categories of AI applications exist, each with unique characteristics.
DEFINITION AND KINDS OF AI
The most common kind of artificial intelligence is narrow AI, which is utilised in a variety of contexts, including big data analytics, mobile devices, and the internet. Because of its capacity to focus on certain tasks and complete activities that are not assigned to it, it is also referred to as “weak” AI. Narrow AI is produced using current standards and technologies and is a complicated computer program rather than a collection of intangible components. Businesses expect AI to do one thing well as its usage grows, which will boost their bottom line exponentially. Narrow AI is created with the newest technology in a setting where the issue is the focus. This gives AI a specific focus, which makes it a preferred corporate option. Expert systems, spam filtering, and recommendation systems are a few instances of narrow AI. Narrow AI is used by recommendation systems, such Netflix’s “Because you watched…”, YouTube’s “Recommended,” and Twitter’s “Top Tweets first,” to ascertain user preferences and provide tailored experiences.[1] Spam filtering, such as Google’s “Spam Filtering” service, uses narrow AI with natural language processing capabilities to keep inboxes clean. The future of artificial intelligence may be paved with expert systems, which are made up of numerous smaller, more focused AI algorithms. Like human intelligence, these systems include a variety of senses as well as cognitive, logical, and creative functions. The future of artificial intelligence is exemplified by IBM Watson, which blends cognitive and natural language processing capabilities.
AI is just a program that can be trained to do jobs; contrary to common assumption, it is not here to replace humans. Establishing a moral, ethical, and legal framework to regulate AI is crucial as it develops. More human-like traits may emerge as the scale moves from typically lower to higher intelligence, making it challenging for developed nations to regulate AI and related technology. As the future of artificial intelligence is carved out, mostly by the kinds of AI that are currently being discussed, it is imperative to have a forward-thinking mindset.
Narrow artificial intelligence (AI) is a very targeted and effective kind of AI that is utilised in many different applications, such as expert systems, spam filtering, and recommendation systems. These technologies can offer users a more efficient and customised experience, opening the door for AI in the future.[2] AI that can think and act like a human, including perceptual tasks like vision and language processing, as well as cognitive tasks like processing, contextual understanding, thinking, and a more generalised way of thinking overall, is known as artificial general intelligence (AGI). The learning component of adaptive general intelligence (AGI) is unsupervised; thus, AGI can be broad and flexible. However, because there aren’t many tools available to develop it, AGI is still a way off. Human intelligence is still a mystery, but neural networks offer a reliable method of producing the precursors of AGI. Since a general artificial intelligence (AI) must be conscious and not only an algorithm or machine, defining consciousness is essential.
Replicating transfer learning, fostering cooperation and common sense, and determining awareness and mind are some of the difficulties facing broad artificial intelligence. Applying knowledge acquired in one area to another is known as transfer learning, and it is crucial to human functioning. Strong transfer learning skills are necessary for an AGI to prevent retraining. The phrase “artificial super intelligence” (ASI) refers to an AI that significantly surpasses human cognition in every manner. Even though ASI is now only a theory, situations involving it have already been imagined. There is broad agreement among experts in the field that ASI will result from the “Intelligence Explosion,” or exponential expansion of AI algorithms. Recursive self-improvement leads to artificial super intelligence, which requires the idea of intelligence explosion. In artificial intelligence, self-improvement takes the form of neural networks learning from human input. The ability of an AI system to learn from itself at quickly rising intelligence levels is known as recursive self-improvement. To achieve genius-level intelligence, for instance, an AGI operating at the level of ordinary human intellect will learn from itself. Artificial intelligence is developing at a rapid rate, becoming more intelligent than itself at every turn. This keeps increasing rapidly until intellect erupts and a superintelligence is created.
[1] T. Ramappan, Competition Law in India: Policy, Issues & Developments, 2nd end. (New Delhi: Oxford University Press, 2009).
[2]EinerElhauge and Damien Gradin, Competition Law and Economics (Oxford: Hart Publishing, 2007).