Quick AI glossary
Below is a quick (and somewhat oversimplified) primer for some key AI-related terms you will come across in the article. If you are relatively new to the world of AI, give it a quick read before jumping in. If you are somewhat well-versed, feel free to jump ahead to the article. If you are an expert in AI, this article is probably too basic for you. (If that’s the case, “Teach me, Sensei!”)
Artificial intelligence (AI): Artificial intelligence involves the theory and development of computer systems able to perform tasks normally requiring human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages.
Machine learning: Machine learning is an application/subset of artificial intelligence in which computers (machines) learn to perceive and act by using statistical algorithms. Think of it as a process that gives a computer the ability to learn without being specifically programmed – kind of like teaching yourself French if you have access to a lot of tutorials and text books about learning French.
Deep learning: Deep learning is a new (and one of the main) branches of machine learning, that is likely bound to revolutionize AI for the future. It is a set of algorithms for performing automatic learning and represents a learning process that goes beyond standard statistical methods. Stretching the French-learning analogy, think of it as you going beyond just mastering French from instructions in the tutorials or text books; but instead, being the language savant that you are, you are now synthesizing a new hybrid language based on your vast knowledge of French and English. Deep learning can be considered as artificial intelligence that is modeled on the human brain. So, instead of being like a cook strictly following a recipe, it acts more like a chef (or at least like a more adventurous cook) who trains to gain knowledge about cooking, and can use the knowledge to make their own recipes one day. AI makes use of artificial neural networks (see below) for deep learning.
Neural networks: Our brain is made of numerous cells (neurons) and the connections between them help us identify billions of patterns and perform so many complex functions. Neural networks are computing systems or algorithms that are modeled after human brains. In AI, artificial neural networks simulate how the brain processes information to learn to recognize patterns and behaviors, and can then make predictions or take actions based on those patterns.
Natural language processing (NLP): Natural language processing is a field of computer science and AI research dedicated to the development of algorithms and methods that enable computers to understand, generate, and interpret human language. The goal of natural language processing is to develop AI that can respond and converse like a human would (“Yes, I’m not a robot”). NLP algorithms and models are based on linguistic theories; to generate sentences and speech, these algorithms use statistical methods to find patterns in human language. NLP has many applications, including automated translation, information retrieval, text mining, speech recognition, and summarization.
From being a staple of science fiction lore or a novelty, AI has grown in popularity and is already being used in many fields. AI has been frequently mentioned in news of late for the massive potential impact it can have on real-world tasks. The high cost training and accessing AI technology was a barrier to businesses that are not named Google, Microsoft, Amazon, or Facebook, but that is changing quickly. While the research and development of AI are still dominated by internet giants, access to AI technology is no longer exclusive to them. It is now becoming more prevalent in various industries. Machine learning and deep learning have resulted in a number of recent breakthroughs in AI capabilities such as natural language processing, language translation, image recognition, autonomous driving, and much more.
Thus, artificial intelligence has been making its way into our daily lives in a number of ways. It is already used in everyday life by the use of smartphone assistants, such as Siri or Google Assistant (“Alexa, order kitten mittens for Aunt Kathy this Christmas”). Based on the huge strides made in natural language processing, it is possible to use artificial intelligence to talk to customers about the products and services or to answer questions about a business. Chatbots that greet you on websites and reply to your text messages are already being used by numerous businesses. AI is frequently used not just to automate human activities, but also for anticipating human needs based on previous behavior. For example, AI can predict your choice of product based on previous experience with your purchases (“but I bought pistachio ice cream from Whole Foods just that one time – why am I still getting it as a product recommendation? Fine, Add to cart it is”). It can recommend movies and news articles that you might be interested in (why does Google keep suggesting you stories about Harry and Meghan vs The Firm? ), and it can make your travel easier (“Yes, Google Maps, I *would* like to avoid the 20-minute slowdown on the highway”). AI personal assistants are being deployed to schedule meetings more efficiently and avoid the endless back-and-forth email chains: (“No, 11:30 am tomorrow doesn’t work for me, but I’m free Friday 2 pm – wait, this calendar invite is for 2 pm Eastern Time, I meant Ostrich Standard Time…”). From these examples, you can see that AI can be used to help with everyday tasks to make them more efficient or fully automate them.
AI can also be used for more complex tasks usually carried out by humans. One such AI-powered technology includes self-driving cars (although impressive advances have been made, it still has ways to go, as demonstrated by this incident). It can also be used to understand the intent of a question using a combination of machine learning and natural language processing. For instance, using a large dataset of question-answer pairs (e.g. Quora or Yahoo! Answers dataset) and a set of candidate answers for each question, the model can be trained to predict which answer is the best for a given question.
Currently, artificial intelligence technology is being used to reduce the time it takes to complete a specific task, and to increase the accuracy of the outcome. In the immediate future, you will most likely see AI tools used to optimize daily tasks that go beyond the scope of its human counterpart. AI is already able to match and surpass the ability of human translators. During this decade, we can expect to see more usage of artificial intelligence trends to take over many basic and some of the more complex manual tasks. This would be supported by data gathering and processing efforts, and combined with human insight and programming.
There is more to AI than just self-driving cars and virtual personal assistants. Below, I will highlight a few areas crucial to daily life where the biggest AI trends are likely to make a massive impact in the immediate future:
AI in healthcare/medicine
The future of AI in medicine is evolving rapidly. More and more people are getting aware of the benefits of artificial intelligence in the medical field. AI-powered tools are increasingly being used in medicine, from pattern recognition to big data analysis. The top ways in which artificial intelligence trends will affect the medical field include (but are not limited to) diagnosis, data gathering, and intelligent automation of medical interventions and treatment.
Diagnosis is a key area that AI-powered technology has the potential to transform dramatically in the near future, given AI has already been shown to be as good as healthcare professionals in diagnosing diseases from eye diseases to cancer. However, as we live in the world where information is at our fingertips, misuse is indeed possible. Instead of self-diagnosing ourselves with mysterious illnesses on WebMD (“it says here that you have internet connectivity problems”), we may go to AI-based mobile apps and get a potentially incorrect diagnosis or wrong medical advice that may have far-reaching repercussions (like this example).
The power of AI in medicine is not limited to diagnosis – AI-powered tools that assist in surgery real-time and also help surgeons train using augmented reality technology already exist. Based on the breakneck pace at which technology trends are evolving, it is predicted that AI will indeed be capable of doing at least some simple surgical procedures like stitching up wounds before the end of this decade.
AI is also already helping researchers predict epidemics, monitor the spread of disease in real time, and develop faster and more effective treatments. It is already enabling researchers to identify therapeutic targets and drug candidates at a lightning pace. Currently, it takes over a decade and a billion dollars to bring a drug from the lab to the patient – think of the time, resources, and lives that can be saved by the intelligent automation or at least acceleration of some of the steps in this process.
The broader adoption of AI in medicine and healthcare in the immediate future is set to transform the way we manage our health, and deepen our understanding of disease and the how the body works. Future research will also explore how it can help meet the demand for access to care and improve the quality of care.
AI in finance
The world of finance offers a wide range of opportunities for intelligent automation, and AI is poised to revolutionize a wide variety of business processes. For instance, it can help analyze trends in data to provide insights into the stock market and predict how the market is going to behave. AI is already finding increasing use in trade execution, helping investors execute stock trades in a fraction of the time it takes humans, and with fewer errors. In the immediate future, AI may even be able to solve complex financial problems without human input, such as the stock market trades and other economic issues. Another example of the use of AI is how financial institutions are now using AI to examine huge amounts of data and identify patterns that allow them to identify more creditworthy customers. This is helping banks and credit unions avoid lending to ‘risky’ borrowers. The issue in such situations, where AI is trained through machine learning models that use historic data, algorithms, and baked-in assumptions based on potential biases is that it can lead to prejudiced forecasts and decisions that can disproportionately affect people who are already at a disadvantage financially (eg, people of color, low-income earners).
AI in media and content creation
The combination of natural language generation/processing and deep learning is a very interesting and promising field. This technology allows us to produce extremely realistic (meaning human-like) text in a wide range of contexts. It can be used to create content for websites, articles, marketing materials, customer service chat bots, and even poetry. As stated before, it has already supercharged areas like speech recognition and language translation dramatically, where AI-powered technology is already as good as or better than their human counterparts and is on the cusp of seeing widespread adoption. It will also enable us to make content creation and curation much more seamless.
AI-powered video generation is a related new artificial intelligence trend where the AI produces video content that is similar to something created by a human. It essentially imitates the natural human ability to create a video based on a script. They can produce large number of videos on many different kinds of styles and topics in a short amount of time. The AI algorithm gets data from the text or audio of a video and uses these data to create the video. Once the AI understands the context, the message, and the information from a video or a set of videos (this can be done by deep learning and reinforcement learning processes), it can create similar videos based on this training. In a society already plagued with the rise of fake news and alternative facts, the power of content creation wielded by AI is bound to present unique challenges.
The ethics of AI
As you can imagine, using artificial intelligent applications and machine learning models for visual content creation is fraught with possibilities of misuse (“Wait, is that a video of Jim from Sales dancing at a Nickelback concert? Oh, it’s a deep fake, I see”).
This and previous scenarios listed in the article about the misuse of AI raise questions about the responsible development and use of AI as well as how these ethics will be enforced.
While the technology aspects of AI have sparked a tremendous amount of interest, the ethical aspects of it don’t receive the same amount of attention. From manufacturing to medicine, AI is invading nearly every industry. However, as the field of AI continues to grow, ethical concerns will also arise. While automation can make the workplace more efficient, the fear of AI taking over people’s jobs is a legitimate concern. Sure, there may be new jobs created by the AI, but these opportunities may require vastly different skillsets from the jobs of the human workers it replaces (this can be considered a point in favor of the argument for a universal basic income). Apart from the ethics around intelligent automation of repetitive tasks resulting in the loss of many types of jobs, there are other ethical concerns we touched on in the article, such as human biases influencing AI learning and behavior, video manipulation (eg, “deep fakes“), and bad actors using AI-powered tools to design dangerous bioweapons.
Many of the ethical concerns mentioned above have more to do with how the technology is put to use and by whom, rather than with the technology itself. Further, there is the aspect of AI eventually developing sentience and its own technology, including advanced weapons – the stuff of literally every sci-fi dystopian movie ever made (“Hello, Mr. Anderson”). AI trains and learns through machine learning algorithms, using input provided by humans — and we’re a messy bunch, so there’s no telling how AI will react to our unpredictability. And because everything AI learns changes its behavior, there’s may not necessarily be any going back once it’s “learned” a lesson.
Overall, it is clear that AI ethics is going to be a critical topic in the field moving forward. The key to developing AI that is widely accepted as ethical is to be transparent about its development and intended use, actively involve AI ethics researchers in the development and decision making process, and educate the public adequately about AI so that they know how to weigh in. It will be increasingly important for the public to understand how AI systems make decisions, what factors go into those decisions, and what can be done to improve the process.
Artificial intelligence: closing words
AI is changing the way we work, live, and interact with each other. It is also changing the way that we think about ourselves and our place in the world. As we try to understand how AI will impact our society, it will also be important to keep in mind the ethical issues that could arise as a result of the technology. Buckle up for the future, because it’s already here.
Hi there! I’m Aaram, the founder of Sciencera. I grew up in the beautiful city of Thiruvananthapuram in the Southern part of India, famous for its pristine beaches. Now, I am a scientific writer based in Indianapolis. When I am not busy procrastinating on my writing, poring through research articles, or coming up with grand ideas to save the planet, I love playing soccer and chess. I read a bit and write sporadically when caffeinated to the right amount. I am passionate about scientific research, writing, and outreach activities.
1 thought on “Top Artificial Intelligence (AI) Trends and How They Will Change the World”