How Artificial Intelligence Is Used Today

How Artificial Intelligence Is Used Today

How Artificial Intelligence Is Used Today

Whether an individual recognizes it or not, they have most likely encountered some form of artificial intelligence. If a person searches on their phone, artificial intelligence is being utilized by the search browser. The user is given the most relevant results based on their past search history, trending results, purchase history, location, and other factors. All of that contributes to returning certain search results to a browser. The following helps describe some of the common areas of artificial intelligence, along with a short description.

Face Recognition

Humans have been recognized by their faces since the beginning of time, but computers are now being trained to recognize faces. Facial recognition has been implemented across many different platforms. From security cameras to mobile phones, facial recognition is widely used. There are various methods by which companies identify and match faces, but Principal Components Analysis (PCA) is one of the most popular [5].

Voice Assistants

Whether it is built into an individual’s phone or connected to a home speaker, voice assistants provide quick answers to questions, help turn the lights on or off, or perform speech translation. The majority of voice assistants can be awakened by a keyword or phrase [6]. Once awoken, the voice assistants can be told a command. What can be done varies from performing a web query to turning the lights off in a house. If a user wishes to turn the lights on in their house, they may use something like, “Hey SMARTDEVICE, turn on the kitchen lights.” The individual’s phone would send a notification to the light bulb manufacturer server, and the server would then send a signal to the light telling it to turn on. There are many smart devices that connect to the cloud, but there are still few devices that make use of local processing [7]. Security is a major concern as most smart devices send sensitive information across the internet. There is the possibility that if someone is able to hack a house’s Wi-Fi signal, they will be able to control various elements of the house including thermostats, lights, TVs, coffee makers, and speakers.

Self-Driving Cars

Cars have begun to offer more self-driving features. Most new cars come standard equipped with lane-drift detection, smart cruise control, and the option to add more features. One of the current leaders in self-driving cars is Tesla. Tesla cars have the ability to drive without input from the user while on highways, and more features are coming to allow them to navigate almost fully autonomously. Self-driving cars work by taking input from sensors and cameras. Based on what data is received, the car’s computer system performs various calculations like how far away the car in front is, which lane the car is in, current speed, and the speed limit. Autonomous cars are one of the heated topics concerning AI ethics. Stilgoe states: “With machine learning, as with other emerging technologies, society has not yet worked out the terms of responsibility, the distribution of liability, the thresholds of acceptable safety or the lines dividing recklessness from negligence” [8]. There have been multiple deaths surrounding autonomous driving, but the question to ask is, Who should be held responsible for the accident? Is the individual within the car responsible? Since the cars utilize machine learning, the algorithms used by the cars’ computer systems will become better over time, but there will undoubtedly be a cost for that learning, including accidents and fatalities.

Individuals enjoy spending time on their cell phones and other connected devices. Things like Instagram, YouTube, Snapchat, LinkedIn, and Facebook can keep people distracted from life. The term “phubbing” has been used to describe the action of snubbing other in-person individuals in order to be on one’s electronic device [9]. Most of the social media services are free, including Instagram, Facebook, Snapchat, LinkedIn, YouTube, and TikTok, while some include the ability to pay for premium features. These services are not truly free; in fact, quite the opposite. Instead of paying for the services with money, individuals trade their privacy for the use of these applications. Companies create a “model” of users. In that model, they store the individual’s interests, which are evident by how much time an individual spends looking at a post or product. The tech giants then display ads based on what the individual’s stored model is perceived to like. “The primary focus of social media, gaming, and most of the Internet in this ‘surveillance economy’ is to gain, maintain, and direct attention—and thus data supply” [10]. Users form almost an addiction to reloading their social media feed. By creating and feeding this addiction, the companies can keep users engaged more often and present more ads. Artificial intelligence provides a way for the digital copies of users to be categorized by what they like and their interests to present more relevant ads and results. In exchange for these free services, users have given up their privacy. These users are often unaware of how well the algorithms know them, possibly even more than they know themselves. Yuval Noah Harari presented the following question: ‘What will happen to society, politics, and daily life when non-conscious but highly intelligent algorithms know us better than we know ourselves?’ [11]. By knowing so much about each individual, algorithms can target users and influence their decisions [10]. This plays a major role in purchases, political viewpoints, and education.

User Recommendations

When a user goes to Netflix or Amazon Prime to watch a movie, they are sometimes overwhelmed by the many options. They can watch anything from a comedy to a horror movie, but it would take forever to watch the trailer of every movie before deciding which one to watch. Netflix uses an AI-powered recommendation system. There are many companies that use a similar approach to recommending items to users. The system in a generalized approach works as follows. Once a user buys/watches/rents an item, they are oftentimes asked to leave a review. On the front end, all the user sees is that the average rating for a particular item might have increased or decreased based on the rating they assigned. But on the back end, companies use that rating to determine what other items that user might like. This is known as collaborative filtering. “Collaborative filtering (CF) is a subfield of machine learning that aims at creating algorithms to predict user preferences based on known user ratings or user behavior in the selection/purchasing of items.” [12]. For example, say that User A and User B rate Movie1, Movie2, and Movie3 as five stars. User A then watches another movie, Movie4. User A gives Movie4 another five-star rating. When User B gets on to their streaming service, they see Movie4 in the recommended movies list. Since User A and User B both previously liked three of the same movies, it is suggested that User B may like to watch Movie4. The relationship between User A and User B can be seen in Figure 2.

Figure 2 – Relationship between User A and User B

As more and more ratings are given, the data model becomes larger. As the data model becomes larger and the algorithm becomes more precise, users will become highly influenced and trust what to watch based on their recommended results since the previous recommendations have been enjoyable.

AI and Ethics

The five examples of artificial intelligence above, face recognition, voice assistants, selfdriving cars, social media, and user recommendations, are all relevant examples of how AI is being used today. Although there are many other uses of AI, the five above will constitute the main discussion of the ethics behind AI. Each presents its own problems regarding security, privacy, moral behavior, and of course, ethics.