Today, most progress in the field of artificial intelligence is completed by the leading companies of the tech-industry. While it is not impossible, it is unlikely that a single person or smaller organization would be capable of producing systems that could compete with larger companies. Even though it may be difficult to further develop these technologies, anyone can use them to some degree. Ranging from Microsoft to Amazon, all the largest tech-companies produce AI based resources for the population to utilize.

 

Penn State AI – Photo Credit

 

Even here at Penn State, these AI resources are being used to some degree. To my knowledge, there are no organizations seeking to develop their own AI based systems, but there are a couple groups that are learning to use and building on preexisting technologies. DevPSU and Nittany AI are the only two Penn State groups that I am aware of. Instead of attempting to create systems similar to those developed by the tech-giants, they use the resources of these larger organizations to develop their own applications.

 

DevPSU – Photo Credit

 

When comparing the two Penn State organizations, it quickly becomes apparent that DevPSU is more beginner friendly. Thanks to their AI-focused learning program, people with little to no coding experience can learn about AI. The program is broken up into weekly installments that last nearly the entire semester. The first few weeks defined artificial intelligence and outlined its many potentials. This involved exploring its history as well as looking into where the field is going.

 

Coming with the educational sessions are several AI-related assignments. These assignments are extremely well-documented and serve as tutorials to using different AI resources, such as Microsoft’s facial recognition system and Amazon’s chatbot framework. Upon completing the learning program, students are then able to participate in DevPSU’s largescale development program, referred to as just DevPSU. The program creates small groups of students and tasks them with projects. Some of them are brought in from companies, some are Penn State oriented, and a few are AI- related. During the course of this program, the teams work together to develop their own project, often using existing frameworks and tools.

 

Nittany AI Challenge – Photo Credit

 

A couple of the AI-based DevPSU teams are actually taking part in the other AI-focused group, Nittany AI. This group is less of an organization and more of a gigantic competition. Nittany AI is an artificial intelligence competition with the focus of creating a project that will help Penn State in some manner. The competition begins with Penn State students forming small groups and submitting a proposal of a possible idea and development plan for a project that would better Penn State. If accepted, the groups would develop their projects. Additionally, the program has several stages, where a number of groups are eliminated, and the successful groups are awarded stipends. The program then culminates with the top 2 or 3 groups getting 30,000 dollars to further their work. These top projects are then often used by the university.

 

While the two Penn State organization that deal with AI are not completing cutting-edge research, they are still creating meaningful projects. Penn State and its students are making a difference in their own community as well as the field as a whole. Even though it might be difficult for them to develop new AI resources, they are doing what they can. They are creating systems that can use preexisting technologies in a meaningful way.

Today, one of the most prominent fields of study within artificial intelligence is machine learning. As discussed earlier, artificial intelligence is broadly defined as computer systems capable of performing intelligent human-like processes. Also previously mentioned, the greater scope of AI includes both narrow and general systems. For the most part, the technologies being utilized now fall under the classification of narrow AI. This is due to the fact that they are meant to operate in very specific situation and only need to perform predetermined tasks.

 

Machine Learning – Photo Credit

 

Meanwhile, machine learning systems are much closer to the classification of general intelligence. This is due to the fact that narrow AI is only able to learn and develop in a certain area, while machine learning is defined “as an application of artificial intelligence that provides systems the ability to automatically learn and improve from experience without being explicitly programmed.” They may not be quite worthy of the classification general AI, but they are certainly closer then the narrow AI systems be implemented today.

 

Essentially, the goal of machine learning is to develop systems that can learn like humans. This takes it a step further than just completing human-like process. Narrow AI systems have expectations and predefined code to produce specific output and receive certain input, meaning only the values are changing. Most AI systems fall into this category as they are built with sufficient base code that is relevant to the tasks being asked of it. Machine learning take it a step further as these systems are not prompted to receive the input and commands, and they are not prepared for the outputs they will produce.

 

It is the differentiation of how the different systems came to be that separates them. Most AI systems are usually classified as narrow and have significant base code that is relevant to the tasks being asked of it, meaning there are already components that take input and produce output. Within a narrow system the algorithm is only modified to make the outputs increasingly more accurate as the system is trained. With machine learning lacked, all of these processes are developed over time as they originate with nothing but the ability to learn.

 

Machine Learning Web – Photo Credit

 

Well then, how do these systems learn? Simply put, it is fed a significant amount of data and creates a model from this data. So, when it is first fed information it finds patterns within the data and creates a model. Once that model is created it attempts to make predictions based on the model based on inputs. This output is then put through a learner which assesses the accuracy of the output by some manner. The learner can be either supervised or unsupervised, which won’t be touched on today, but is in the graphic above. Once this is done, the parameters of the model are modified to make the model more accurate as a whole so that it will be able to make better predictions in the future. This system differs from narrow AI systems in that the basics of the model would be initially provided, giving the technology the fundamentals before training it.

 

Machine Learning Explanation by Google – Photo Credit

 

Machine learning may seem unnecessary when looking at it compared to other forms of artificial intelligence that have frameworks related to what they are doing. However, by setting these systems free with no basis they can be applied on larger scales and one system can be implemented in a variety of ways. Today, machine learning is being used for visual recognition, speech recognition, medical diagnoses, statistical predictions and much more. It is difficult to determine the limits of this technology today and it will be interesting to say what it will accomplish.

One of the most astonishing effects of artificial intelligence is its potential impact on our economy. Permitted that we allow artificial intelligence to infuse itself with our many industries, they will grow at an astronomical rate. Accenture, a global research and consulting firm that offers much more, has investigated and published their findings of how significant AI’s impact on the global economy will be. These statistics are based on current systems in each industry and speculates what roles AI might have as well as the prospective economic gain.

 

Firstly, only twelve countries are developed enough to harness AI in a meaningful manner. Accenture equates this to utilizing AI in a manner that significantly boosts labor productivity by up to 40 percent. Ranked in order of impact on labor force, the twelve countries are Sweden, Finland, United States, Japan, Austria, Germany, Netherlands, United Kingdom, France, Belgium, Italy, and Spain. These countries currently have the ability to spread artificial intelligence to the rest of their economic infrastructure in some manner or another. Essentially, this means that artificial intelligence will have an effect most of their industries to the point that it will be able to significantly boost the productivity of labor.

 

Impact on Labor Productivity – Photo Credit

 

Thanks to the artificial intelligence’s ability to boost labor productivity, these countries are expecting to roughly double their annual economic growth rates. Some countries will just experience a little less then a double, but Japan will over triple their current economic growth rates. This will in turn allow these twelve countries to double the size of their economy much faster. According to this 2016 study, the United States will be able to double the size of their economy by 2040 thanks to AI. Without artificial intelligence, it will take until roughly 2060. One of the most shocking cases of this is Japan. By implementing AI, they will be able to double the size of their economy by around 2055. Without AI, it is projected to take until around 2115. Clearly, the impact on the global economy will be sizable.

 

Time to Double Economy – Photo Credit

 

In the past, AI was primarily thought of as a means for job automation. This would primarily involve having computers replace humans in certain positions. Ideally, they would be able to fulfill these roles more efficiently and effectively. This would be artificial intelligence’s primary manner of impacting the labor force, by adding to it. However, there has been a shift more recently. It has been found that they can have an impact far beyond automation. In the future, AI will be set to complete all sorts of data analytic roles to find means of taking on more profit. Additionally, there will be many AI-based services and products that will be able to assist users far better than normal algorithms.

 

Profit Growth by Industry – Photo Credit

 

As me move forward, AI’s impact on our global economy will be undeniable. It is set to significantly increase the profits of several industries. One of the most notable is education, which is projected to be receiving 84 percent more profit in the year 2035 if AI is utilized. That is just one of the many cases in which AI can have an impact on our many industries. If correctly implemented on a large scale, artificial intelligence will have a humongous impact on our global economy moving forward.

I’m hoping by now that I’ve managed to express the theme that artificial intelligence has many applications. Of the many emerging AI applications, cybersecurity is one of the most prominent. For those of you that don’t know, cybersecurity is defined as “the practice of defending computers, servers, mobile devices, electronic systems, networks, and data from malicious attacks.” Simply put, the aim is to make sure that all of these systems are secure.

 

Cybersecurity – Photo Credit

 

These systems need to be made as secure as possible to ensure the overall safety of everyone. This is due to the fact that they all house sensitive information. Ranging from your own passwords to the president’s nuclear launch codes, sensitive data exists everywhere. For the most part, this information is marketable, leading many criminals to try and access it. Because the majority of systems are connected in some manner or another there are pathways from one computer system to another. If these are not properly secured they can be exploited, allowing important data to be harvested by the wrong people. Cyber-criminals can then sell or use what the harvest. Depending on the situation this can have terrible implications for everyone.

 

Cyber-Attack – Photo Credit

 

Cyber-threats can take on many different forms. In addition to stealing data, some cyber-attacks involve taking down systems. I know the company I worked for encountered an issue along those lines last year. Essentially, someone was repeatedly making requests to their servers in an attempt to block anyone else from using it. This prevents their customs from being able to make purchases, which then leads to a loss of sales. Whoever was doing this then asked for the company to pay a ransom to make them stop. Instead they refused, increased their security, and are no longer experiencing the problem. This was just one instance of a cyber-threat. They exist everywhere. Who knows, you probably have a couple sitting in your email right now, waiting for you to give them access to your device or accounts.

 

As technology develops, it becomes a more powerful tool. This means that it can have many new positive effects on society. However like most things, it can be applied in a negative manner. For this reason, there is a constant need to continually update cybersecurity. The United States Government is a prime example. They spend 19 billion dollars a year to maintain and develop more effect cybersecurity systems. Of many developments in this field, the introduction of artificial intelligence is one of the most revolutionary. The goal is to develop AI-based systems that can detect threats significantly faster and more accurately than humans. However, it comes with some risk, especially early on.

 

AI Cybersecurity System – Photo Credit

 

Using artificial intelligence will allow for the systems to develop overtime.  After lots of practice, AI-based systems will be able to more accurately distinguish threats, even as the threats change overtime. Because it will be based on one system, the underlying system will be vulnerable if it fails. For this reason, a system built of many diverse algorithms will be the most successful. When this is achieved, it will limit the required work of the employees and will hopefully still remain more successful.

Many of you have heard about self-driving cars in some way or another. However, I doubt the media has taken the time to explain how they work. Simply put, AI are how these self-driving cars work. Yes, computers are already driving test cars and they will be operating these vehicles in the future.

 

Self-Driving Car – Photo Credit

 

First of all, let’s address why self-driving cars are being developed. At the center of the movement is the issue that humans are just awful at driving cars. A recent study reported by IoT found that “over 90 percent of road accidents are caused by human error.” This includes speeding, drunk driving, distracted driving, and much more. Autonomous have the potential to be much safer because they will continue to operate at the same optimal level no matter what. Self-driving cars will never be impacted by human hindrances like illness, fatigue, or any mental or physical impairment. Lastly, self-driving cars will always be more attentive to their surroundings than humans. Hopefully, self-driving cars can become safer than today’s norm.

 

High Percentage of Driver Error – Photo Credit

 

As technology develops, so do the possible applications. Years ago, self-driving cars would not be feasible. Now, this is possible thanks to three different technologies. Sensors, connectivity, and advanced algorithms are all utilized to create these self-driving cars. Within the cars are many, many sensors. Essentially, they entire vehicle is covered in sensors. This allows the system to detect everything. Cameras, sonar, radar, and other sensors allow the car to determine everything around it at once. On top of all this information is the connectivity of these cars with larger systems, such as the cloud, allowing them to take in the data that their sensors can not pick up. This includes traffic patterns, weather maps, road conditions, and more. In the future self-driving cars may be able to communicate with each other to determine the intentions of the vehicles around them. Clearly, self-driving cars will be able to take in much more information than we humans could ever receive at once. But, what will they do with it all?

 

Self-Driving Car Sensors – Photo Credit

 

This is where the AI comes in. The aforementioned advanced algorithms include artificial intelligence. Because computers are able to process much more information than us, they are able to receive the piles and piles of data and translate that information into concrete commands. The goal is for these commands to one day become flawless, seeing as we clearly are not there yet. AI are used in these cars, because it allows for the algorithm to become better and better over time. Artificial Intelligence involves the systems learning on their own. This does not mean that it will become self-aware and rise up against us. I mean that it is capable of modifying its actions slightly after testing and finding out what works best. So, these self-driving cars will become better and better with practice as a human would. They need to learn what to do in any given situation and how to best apply this knowledge.

 

As of today, true self-driving cars have not yet reached the market. There have been cars released with self-driving modes, but they are very limited. The cars are only able to use these systems in certain situations and the human is still required to maintain awareness, so they can step in if necessary. Of course, this is a step in the right direction, but the current systems do not reflect the true potential. There are complications in that several people have been in accidents in these vehicles. However, human error was the cause of the majority of these cases. Still, these cars cannot react to ambiguous situations yet, such as a drunk driver swerving towards them. Before they can be applied on a large scale, they must be trained to deal with all situations, not just the limited trials put forth by their creators. While they might not be quite there yet, self-driving cars have the potential to make the roadways safer.

Today Chatbots are one of the most prominent examples of AI. Whether you realize it or not, it is more than likely that you have come into contact with a Chatbot. Many social media applications have integrated Chatbots to assist their users. These can be artificially intelligent, but they are usually rule-based, meaning that they can only respond to specific commands. When these distinct commands are inputted by the user, the system is able to produce the correct output depending on the system.

 

Chatbot Example Graphic – Photo Credit

 

For the purposes of this blog, we will instead focus on artificial intelligence based Chatbots. These can be integrated with larger platforms or exist separately. I recall that my first interaction with a Chatbot was on Kik. This was roughly 6 years ago. I was riding on the bus home from school when I got a notification from a stranger with a stock photo. I was told not to talk to strangers, but what can I say? I was curious. The bot started the conversation by saying something generic like “Hello.” It asked several random questions and attempted to talk with me as though it was an actual person. However, it was quite obvious that it was not a real person. I would ask questions or answer their inquiries completely wrong and it would not adjust its outputs. As far as Chatbots go it was pretty terrible, but I must commend its creator for their efforts.

 

Now, Chatbots have significantly developed. Many larger organizations have developed artificial intelligence based Chatbots that are able to take dynamic user inputs and generate unique outputs per the request. These Chatbots work by first tokenizing the user inputs. Language is quite weird and can be ordered and interpreted in many ways. For this reason, a natural language processor is used to break up the input into values that a computer can reorganize and understand. This even involves recognizing key words like names and normalizing the text, by correcting grammar and spelling so that it can be understood. Once this is done parts of speech are assigned and the input is parsed into simpler phrases that can be understood. Once this is completed, the program can output the relevant output to the user.

 

Depending on the purpose of the system, the output can be very different. Some of the most sophisticated systems are already within our reach. Siri, Alexa, Cortana, and Google Assistant are examples of virtual assistants that utilize Chatbots to better serve their purposes. The Chatbots integrated in these systems allow them to understand what the user is trying to communicate. The other aspects of their systems allow them to calculate and find the appropriate output. However in the majority of these cases, the Chatbots are only acting as assistants and fulfilling minor requests. The conversation is there but limited.

 

Cortana, Microsoft’s Personal Assistant – Photo Credit

 

While it is nice to have these personal assistants in our pockets, Chatbots can be used for far more. One new development is that Chatbots can be used as a companion for the elderly, especially for those with Dementia. It will provide these individuals with someone to talk to, which can be especially important if they are alone. Endurance is one of these systems and it is specifically designed to help people with Dementia. It does this by providing them with help as they struggle with their daily routine. Similarly, Chatbots can be used to help people that are unable to sleep. Casper is a new program that is meant to help get insomniacs through the night. Essentially, it provides people with someone to talk to when no one else is awake, or around.

 

Casper Chatbot – Photo Credit

 

It is important to understand that Chatbots can be used for more than the applications I outlined above. They are already being used for much more. The list ranges from generating faster medical diagnoses to pure entertainment. I have even heard of creating a suicide prevention system. The list goes on and on. Moving forward, I am certain that Chatbots will be implemented in many different ways and they will only become more prominent.

I’m certain that many of you have heard of Artificial Intelligence. But, what really is it? AI has been featured in many movies, such as Ex Machina and The Terminator series. Typically, these movies depict AI as human-like robots that are completely self-aware and capable of destroying the entire world. In many ways, these depictions fail to properly display artificial intelligence.

 

The Terminator – Photo Credit

 

For one thing, it is extremely difficult to definitively determine what is AI and what is not. The Merriam Webster Dictionary provides a very simple and broad definition of artificial intelligence, stating that it is “a branch of computer science dealing with the simulation of intelligent behavior in computers.” Falling under this umbrella is anything that we, as humans, define as an intelligent. This status is reserved for software that is extremely clever, even human-like. As you can see, this leaves a lot of room for interpretation.

 

In an attempt to categorize the many types of artificial intelligence, they can be categorized as either Weak or Strong AI. Weak AI is also known as Narrow AI. This is because it focuses on a single task and is developed to a point in which it can perform this task better than humans. When combined with many other Weak AI systems, Strong AI can be created. This second category of AI refers to the systems that are usually depicted in the movies. Strong AI refers to machines being able to think and perform tasks on their own, similar to a human. For example, it would be able to complete a task of your selection which it has not been trained to complete. It would behave and learn in a way that is similar to a human. It would continue learning to complete tasks, but it is unknown whether anything that is truly self-aware can be created.  So, I wouldn’t worry about computers taking over the world just yet.

 

Strong vs Weak AI – Photo Credit

 

At this point in time, there is no real example of strong AI. However, there are already many examples of Narrow AI. These systems are only developed to complete single tasks and are already being integrated with society. Because they are task oriented, they cannot cause any harm on their own. One way of thinking of it would be that Weak AI is just better software. Instead of just creating a program that completes a certain task the same way every time and continues to perform at the same level for the entirety of its existence, Weak AI can be used to create a system that will continue to improve. This is because the Narrow AI is set up to receive feedback based on its performance and continue to improve as it encounters more practice. It cannot develop new functions, just improve.

 

As mentioned earlier, this type of AI is already being integrated within society. Because of the way it is set up, it can perform certain tasks much better than humans and continue to get better with more experience.  You might think that his would only work in certain areas, but its applications are virtually limitless. From chatbots to facial recognition, AI is set to help humanity perform certain tasks that it wouldn’t be able to do on its own. There is no saying what AI will accomplish, but I can say with complete confidence that its impact will be momentous.

 

Potential Economic Impact – Photo Credit