Function-based AI
The criterion used to determine the type of AI depends on how a machine compares to humans in terms of versatitilty and performance. An AI that can perform more human-like functions with equivalent levels of proficiency is a more evolved type of AI.
Function-based AI
Concerns
Applications
Implementing AI
Capability-based AI
Capability-based AI
Introduction
Narrow AI
General AI
Application
Super AI
A reactive machine only works with present data—it does not store memories or use past experiences to determine future actions. Reactive machines perform specific tasks and do not have any capabilities beyond those tasks. Reactive machines always respond to identical situation in the exact same way every time like the recommendations that Netflix makes based on your past viewing.
Ability
to
Think
Ability
to
Feel
Reactive Machines
Limited Memory
Pedestrian
Status: Walking
Direction: Crossing
Alert: Elder
Cyclist
Status: Moving, 5 mph
Direction: Crossing
Limited Memory AI uses past data to make decisions. Self-driving cars are Limited Memory AI. They use data collected in the recent past to make immediate decisions and make adjustments as needed. This information is not saved in the car’s long-term memory though.
Traffic Light
Status: Red
Direction: Main
Bus
Status: Moving, 25 mph
Direction: Crossing
Car
Status: Stopped
Direction: Main
Theory of Mind AI covers systems that are able to understand human emotions and how this affects decision making. They are trained to adjust their behavior accordingly.
Theory of Mind
I am so sorry that you are frustrated
with this route. I will remember not
to take this way in the morning.
Self-awareness
Self-awareness systems will need to understand their internal traits, states, and conditions in addition to perceiving human emotions. These machines will be smarter than humans and will be able to undersand and evoke emotions in those they interact with. These systems will have emotions, needs, and beliefs of their own.
Learning
the process of acquiring new understanding, knowledge, behaviors, skills, values, attitudes, and preferences
AI is the ability of a machine to display human-like capabilities. It enables a technical system to perceive its envionment, deal with what it perceives, solve problems and act to achieve a specific goal.
Planning
the process of thinking regarding the activities required to achieve a desired goal based on foresight
click on each iconAll technologies are categorized by their
capacity to mimic human characteristics.
Reasoning
the action of thinking about something
in a logical, sensible way
Creativity
the use of the imagination or original ideas
Narrow AI is also known as Weak AI, and focuses on one narrow task. Narrow AI targets specific tasks that it is programmed to do and has problems with tasks outside its programmed abilities.
Facial Recognition
Autonomous
Car
Voice Recognition
General AI, also known as Strong AI, can understand and learn any intellectual task that a human can. Machines apply knowledge and skills in various contexts. Researchers are still working to achieve strong AI by trying to find a method to make machines conscious and to program a full cognitive ability set.
General AI
Super AI
Super AI surpasses human intelligence and can perform any task better than a human. Some of the critical characteristics of super AI include thinking, solving puzzles, making judgements, and decisions. The existence of super AI is still hypothetical.
Implementing AI
Concerns
Application
Implementing AI
Transportation
Sensors, AI and machine learning technologies, and big data analysis all play a role in autonomous driving. Vehicles today include sophisticated advanced driver assistance systems (collision warning, lane departure, traffic signal recognition) that when combined with machine learning to process the data from cameras or radar make autonomous driving safer. AI technologies also manage traffic, predict flight delays and make ocean shipping more efficient.
Manufacturing
Industrial robots have evolved from performing single tasks separate from human workers to multitasking robots that collaborate with humans in warehouses, factory floors and other workspaces.
Applications
Law
AI helps to automate labor-intensive processes such as the discovery process which requires the review of many documents. Law firms are using machine learning to describe data and predict outcomes.
Healthcare
Machine learning helps to make better and faster diagnoses. Online virtual health assistants and chatbots help patients find medical information, schedule appointments, and understand billing or administrative processes..
Finance
AI software performs much of the trading on Wall Street. Other applications collect personal data and provide financial advice and can even help in the purchase of a home.
Artificial intelligence has become an integral part within many markets.
<
Business
Customer relationship managment (CRM) platforms integrate machine learning algorithms into analytics to better serve customers. Chatbots provide immediate answers to customers when incorporated into company websites.
Education
AI can automare grading. AI tutors provide additonal support to students to help them stay on track.
Lack of
Regulations
While the impact of artificial intelligence on society can be very beneficial, there are several concerns that need to be examined.
Automation makes it more difficult to detect evil acts like embedding viruses into software and manipulating AI systems for personal gain.
Security
Problems
AI
Bias
Some industries operate with strict regulatory compliance requirements and the way that a decision was made needs to be known. When AI tools make a decision, they look at correlations between thousands of vaiables so there is not an exact reason for a decision. Black box AI occurs when the decision-making process cannot be explained.
Imagine the development of autonomous drones and robotic swarms, remote attacks, diseases delivered by nonorobots. or computers overtaking humans altogether.
Machine learning algorithms are only as smart as the data they are given in training. Since a human selects the data used to train an AI program, there is the potential for machine learning bias, results that are systematically prejudiced due to erroneous assumptions in the machine learning process.
Need for
Data
The data needed for AI programs must be accurate and pure. A great deal of effort and money goes into making data AI-ready.
Personal
Touch
Concerns
Terrorism
Machines cannot replace the interpersonal relationships that strengthen a team. Additionally, AI is constrained by rules and algorithms and cannot exhibit the creativity of humans.
Laws
Big Tech
Dominance
Big Tech firms dominate search, social media, online shopping, and app stores and are the main providers of AI to the rest of the market.
Laws that regulate AI will not be easy because AI comprises a number of technologies. The rapid evolution of AI technologies is another obstacle to meaningful regulation.