There’s a learning curve that comes with operating as a citizen of the internet. Browsing different websites and liking videos is second nature at this point, but the policies and practices driving that content engagement process is murky at best. We’re not all IT professionals and computer engineers, but we shouldn’t have to be to understand the basic processes driving the content we consume everyday.
With that in mind, our team sat down and identified the top terms in data and AI that you need to know:
If you spend time on social media, then you’ve inevitably read posts decrying the unpredictable whims of that platform’s algorithm. At its core, an algorithm is a formula built to solve a problem under specific conditions. Algorithms take a current state and adjust key variables to create a desired output.
Online, a social media algorithm prioritizes the content you see based on what is most relevant to your interests. Where the early days of social media populated your homepage in chronological order, social media algorithms use your previous behavior to identify the people, mediums, topics, and even content lengths that you are most likely to engage with. In this circumstance, the algorithm’s task is to keep your attention on the platform by providing desirable content and filtering away low-effort, irrelevant, boring, or offensive material.
While most organizations commit countless work hours to optimizing their algorithms to reach the perfect content mix, at Invisibly, our users control their algorithm. Instead of a machine guessing, you can optimize your own experience — and earn money in the process.
Algorithms are possible because of machine learning. Machine learning works under the principle that different systems—algorithms, programs, artificial intelligence, etc.—can identify patterns in large data sets and use those patterns to make informed decisions without help from a human operator. For example, a social media algorithm might monitor your engagement and determine you’re 50% more likely to watch a cat video than a make-up tutorial; email providers use machine learning to block spam and filter out potentially dangerous messages. Machine learning constantly empowers real-time adaptation in the tools and devices we use every day. So, when you start your commute in the morning, and your phone asks if you’d like to open a podcasting app—that’s machine learning.
Machine learning thrives by consuming huge data sets from multiple sources. For example, Netflix uses data from its millions of viewers to make generalized recommendations for what viewers with similar interests might enjoy. Then, the streaming service takes that recommendation a step further by using your own habits and preferences to create a specific recommendation that aligns with your unique interest. Those specific variations of one individual give machine learning the power to customize the experience on a person-to-person basis.
Your social media feed is the primary hub where you experience content on a social media platform. Every individual and organization you follow has their posts, reels, and stories funneled into the feed, then processed through the algorithm to deliver the latest content directly to your preferred device.
Different platforms condition their feeds to service the needs of their unique audience. Because different people visit different social media platforms for different purposes, what convinces one person to stay on Instagram would likely run someone off of LinkedIn. Your social media feed is constantly adapting to create a more compelling and customized experience.
Scrolling is the action of moving up, down, or sideways through a website or app to view new and different parts of the platform. Social media and many other digital experiences often include an “infinite” scroll feature that automatically populates the bottom of the page with new content and creates an uninterrupted browsing experience.
The goal of any platform is to keep its users scrolling. As highlighted in The Social Dilemma, social media channels keep you engaged by using the infinite scrolling feature, creating a unique experience catered to your needs and interests.
What is AI?
Artificial intelligence (AI) is the recreation of human cognitive capabilities inside a machine. These machines use deep data sets to simulate human behavior. For example, factories use AI to recreate manufacturing tasks; search engines use AI to provide personal recommendations. To some degree or another, machine learning and algorithms are a function of AI.
The goal of AI is to create a machine that can think logically. In its basic form, developers train that system under a series of “if-then” statements—“If the ball is red, then it gets put into the bucket on the left,” for example—and scale that complexity to meet the needs of the desired action.
What are the different types of AI?
Not all AI is created equal. As that complexity scales, so does the capabilities of the AI itself. To date, AI falls into practical and theoretical categories.
Reactive machines & limited memory systems represent Artificial Narrow Intelligence, and the most common AI we experience on a daily basis.
Theory of mind & Artificial General Intelligence drive hypothetical machines that perfectly recreate human intelligence.
Artificial Super Intelligence & self-awareness elevate the capabilities of the machine beyond that of its human operators into something entirely independent. These machines are entirely hypothetical and met with considerable skepticism among the general public and tech community.
a. Reactive machines
In their simplest form, AIs are reactive machines, limited to the conditions in front of them. Reactive machines don’t learn or adapt but only act according to their purpose and programming. That’s not to imply that all reactive machines are simple; IBM’s Deep Blue supercomputer was a reactive machine programmed to play chess. Reactive machines don’t need to understand any concepts beyond the task in front of them.
b. Limited memory
By comparison, limited memory Ai are Type II machines that look to previous experience to inform future action. Self-driving cars are the most common example of limited memory AI. By combining observational and integrated knowledge, self-driving cars analyze every element of their environment and predict the future action of pedestrians and traffic around them in order to travel safely.
c. Theory of mind
While reactive and limited memory AI are relatively easy to find in modern technology, the next two are purely conceptual as of this writing. Theory of mind is the idea that people act according to their desires. We’re tired, so we sleep. We’re hungry, so we eat. Where reactive and limited memory machines function within their own tasks, theory of mind machines elevate their ability to understand the why behind their actions. For example, a self-driving car needs to know what the road is, what a car is, what a pedestrian is—theory of mind expands those capabilities to where the vehicle not only understands what, but why a driver might swerve into a lane, or a pedestrian may step into the street.
d. Self-awareness
The pinnacle of AI, self-awareness is the state where a machine is developing its own ideas and desires. Self-awareness comes with an understanding of what motivates action. While we as humans can generally speak to why we make certain decisions, machines are traditionally limited to their programming. Self-awareness shifts that decision-making process from “I am acting” to “I know why I’m acting.
e. Artificial Narrow Intelligence (ANI)
Artificial Narrow Intelligence represents AI as we know it today. ANI systems perform a specific task with human-like capabilities and decision making, according to the logic and rules it was taught. This generation of machines is limited to its programming. Even the most complex limited memory systems are anchored in some way to a basic list of instructions.
f. Artificial General Intelligence (AGI)
Artificial General Intelligence is a hypothetical AI system capable of learning, understanding, and acting in the same way a human would. As a result, AGI’s can pursue new skills and generalizations on their own, without the guidance of a human operator. Essentially, AGI represents a machine that thinks, learns, and acts in the same way a human would.
g. Artificial Superintelligence (ASI)
Artificial Superintelligence amplifies the capabilities of a machine beyond mere human capabilities, creating a thinking, learning, decision-making process that not only demonstrates human-level intelligence but executes that logic in an instant.
Access should never be an obstacle. Since the very beginning, we at Invisibly strived to invite everyday people into the conversation around data, AI, and the technologies shaping the future of our online experiences. If you have questions about the key terms and topics driving today’s digital landscape, feel free to message us on social media or visit our Data Terminology: Important concepts and definitions in data blog to learn more about the terms and topics shaping our future.
Access should never be an obstacle. If you have questions about the key terms and topics driving today’s digital landscape, feel free to message us on social media or visit our Data Terminology: Important concepts and definitions in data blog to learn more about the terms and topics shaping our future.
Use your data to access premium content you love.