Artificial intelligence has been a driving force behind the digital transformation that is characterizing the current IT and business landscape. The information age has given us access to more data than humans can process, and relying on AI systems to work with vast data repositories is empowering us to find new ways of deriving value from information.
We don’t have the kind of thinking and feeling machines that science fiction movies portray, but the many uses of Artificial Intelligence are sources of considerable progress for several industries.
A brief history of AI
The concept of AI is not new, but a lot of current applications weren’t possible until computer science and hardware innovations caught up. Here are some significant milestones in the history of AI:
- In his 1950 paper “Computing Machinery and Intelligence,” Alan Turing formulated the idea that humans rely on information and logic systems to make the best decision possible and that this model could be transposed to machines.
- In 1956, the Logic Theorist program organized by John McCarthy and Marvin Minsky brought together researchers from different fields to discuss AI. The term artificial intelligence was coincidentally coined during this conference.
- Technological innovations throughout the 1960s and 1970s were instrumental to the development of AI. Machines could store commands rather than merely executing them, and memory miniaturization made complex machines more cost-effective.
- The 1980s saw the rise of new concepts like algorithms, deep learning and expert systems for decision-making. The Japanese government funded research to develop expert systems, marking one of the first uses of AI in a business setting to duplicate the decision-making process of experts in a given field.
- IBM’s Deep Blue decision-tree AI beat world chess champion Gary Kasparov in 1997, a symbolic milestone for AI outperforming humans.
Different types of AI
Functional AI systems fall into the artificial narrow intelligence, or weak AI, category. These systems operate with predefined functions and are designed to perform a specific type of task.
There are different types of systems within the weak AI category:
- Reactive machine AI. One of the earliest examples of reactive machine systems is IBM’s Deep Blue AI, mentioned above. It makes decisions based on the current configuration of the chess board, and all the possible configurations are built into the initial algorithm. The system doesn’t learn from the outcome of previous games.
- Limited memory systems. Most modern AI systems fall into this category. These systems have the capacities of reactive machine systems, but they can use historical data to make decisions. Self-driving vehicles and personal assistants are based on limited memory systems.
- Theory of mind. This branch of AI aims to develop systems that can incorporate factors like human needs and emotions. It’s still theoretical.
- Self-awareness. In theory, AI systems could achieve capabilities like consciousness, representation of self and internal states.
Strong AI, or artificial general intelligence, replicates the way humans function by drawing parallelisms between different areas of knowledge and handling a much wider range of tasks. These systems are still theoretical but could open up infinite possibilities since they carry out complex tasks without being limited by predefined functions.
Artificial superintelligence is the next step in AI evolution, with capabilities that exceed human thinking.
Top AI trends and branches
There are some innovative concepts shaping the development of AI within the limited memory system category.
Machine learning. Machine learning is a major trend shaping current AI uses. An AI system is fed massive amounts of data and looks for patterns. The more experience a system has, the more accurately it finds patterns. There are different methods used to provide feedback and improve the system, including supervised, unsupervised and reinforcement learning. AI systems like Netflix’s recommendation algorithm use machine learning.
Deep learning and neural networks. Neural networks possess several nodes that act as steps in a complex decision-making process. These AI systems are attuned by attributing more or less weight to a specific node to achieve a more accurate outcome. The purpose is to create a system that can process varying input. Examples of deep learning and neural networks include speech recognition and automated translation tools.
AI applications
Various sectors are embracing AI and driving research.
Health care. Health care is a field that can greatly benefit from adopting new techniques to derive value from medical data. AI is the perfect answer.
IBM’s Watson is an AI platform that sifts through data from clinical trials and matches patients to the most relevant trials. It’s an example of a task that a human couldn’t perform due to the sheer size of the data set. Here are other uses of AI in health care:
- AI systems can draw on historical data from other patients with similar symptoms to formulate a more accurate diagnosis or use technology like image recognition to support medical imaging diagnoses.
- An AI system can adjust medication doses based on a patient’s symptoms, history and biology.
- Natural language processing can sort through medical literature and identify drug interaction risks. Japanese company Sumitomo Dainippon Pharma pioneered a new way of creating treatments with an AI system that generated a molecule for an OCD treatment.
- AI and predictive analytics are valuable tools for gaining insights into how a disease is likely to progress in a patient or how an outbreak could spread.
- AI can power features like chatbot interactions for telehealth patients.
Business processes. AI has revolutionized the corporate world with uses like process automation. Repetitive tasks such as updating customer files, billing or marketing communication can be automated to save time.
We’re also seeing businesses implement AI systems to optimize existing processes. An example is fleet management with an AI that creates the most time- and cost-effective routes.
Analytics solutions leverage AI to derive value from data and provide businesses with actionable insights and predictive analytics that support decision-making.
Security and fraud prevention are other areas where businesses can benefit from adopting AI. Modern fraud prevention systems analyze data from payments and look for unusual patterns. They also monitor traffic in real time and stop attacks.
AI and the IoT. The internet of things is another area where we’re seeing new uses for AI:
- Smart home products rely on AI systems to deliver a more personalized experience and automate actions based on preferences or behaviors.
- Self-driving cars use AI systems to analyze the environment and make the best decisions possible.
- Industry 4.0 needs AI systems to work with the data generated by connected machines and data points across the supply chain.
- We could see more AI-controlled drones and robots in the future.
Getting started with AI
Are you curious about how AI can benefit your business? AI-powered anomaly detection for fraud prevention is a solution that can benefit any business that engages in digital transactions.
Rules-based fraud prevention processes are no longer a match for the kind of sophisticated schemes criminals are putting together. Fraud.net’s solution is to combine the power of deep learning with consortium data from multiple sources to detect unusual transaction patterns in real time. Learn more about the technology we use and get started with a free trial.