For years AI was touted to be the next big technology. Expected to revolutionise the job industry and effectively kill millions of human jobs, it became the poster child for job cuts. Despite this, its adoption has been increasingly better received. To the tech experts this wasn’t really surprising given its vast range of uses. Here is a list of some of the most notable applications:
Your actions on the internet create data. Many companies store that data and feed it into different machine learning models which then power predictive algorithms.
Through this, platforms like Amazon and Netflix give you suggestions on what you may want to buy or watch. This predictive power is not only convenient for users but also to those behind these platforms, who can analyze trends better and run their businesses more efficiently. Reviewing this high volume data manually is impossible and highly prone to error, but automating it will likely help identify more trends, some that might even be too subtle for the human eye.
Virtual Assistants and chatbots
The earliest traces of a chatbot were in the early 1960s. Designed by MIT professor Joseph Weizenbaum, ELIZA was an attempt to demonstrate communications between humans and machines. It was able to pass the Loebner test, one designed to check if a machine could understand an input (of any form). While ELIZA famously never understood what people would say to it, the chatbot could respond convincingly leading to people being persuaded that it was intelligent despite the creator’s disagreement. Since those days of having to type out our side of the conversation and hoping that a relevant answer would return, we have come quite far.
With virtual assistants like Alexa and Siri, they are essentially performing the roles of glorified chatbots, except conversations are a lot more natural and reliable. Depending on the amount of control you give to these devices, they can go from just answering questions about the weather to switching on the lights in your room to reserving an appointment at the dentist. Tools like this bring convenience to your fingertips.
Every step of the way in the healthcare industry can be improved with AI. From disease detection and prevention to minimizing surgical errors to even drug creation. AI has a large role to play and one that will likely raise the average life expectancy, significantly for those who can’t afford current healthcare solutions.
With consistent logging of each visit, a thorough review of the symptoms might reveal a disease that even a seasoned doctor might miss. A deeper knowledge of similar cases around the world could also lead to quicker diagnosis and even warnings of a larger trend- like a pandemic.
As a result of this cataloged data, it may even be easier to run simulations which eventually will lead to more efficient drug creation.
Exploring the great unknown has been an eternal passion for mankind but one that we have simply scratched the surface of. While in the past there have been some limitations be it technical or physiological; we have incrementally worked around the challenges. Now with established satellites and rovers, there is a heavy influx of data. With AI, this can be analyzed and we can begin to understand the world beyond our own better. Not only can AI be used for this but in other scenarios as well: The same underlying technology behind self-driving cars can be adapted to identify asteroids and help identify and solve on-board problems. What’s more, is that AI can make it safer for humans to travel in space or can just replace high-risk travel altogether.
This is just one of the many forms of AI. Providing companionship is at least wholesome use of this technology. Many countries in the world have an aging population and with it a larger group of senior citizen inhabitants. Typically they tend to live alone and have limited social interactions. As a result, they are more prone to developing dementia and depression, over and above their pre-existing medical conditions. With these AI companions, not only do they get a tool to communicate with but one that can monitor their health and even save them if a medical situation arises.
Similarly, these AI companions can be adapted to work with children as well. This doesn’t limit it to just conversing, but also keeping them safe.
Automated gaming created the initial buzz of AI among the public. While people in the realm of technology even in the late 1980s knew the potential of AI, the public was skeptical until IBM’s Deep Blue beat Grandmaster Garry Kasparov in a game of chess in 1998. Through this, the world finally began to grasp the true power of AI and realize how close the world was to developing useful technology with it, something that, until then, only seemed attainable in the distant future.
The reason that even early AI algorithms performed well in gaming was and still is because there is a clear objective and set of rules that all players must function under. There are minimal deviation and only a limited set of steps that can be played. Since most of these computers have exponentially larger computing power than humans, it can simulate a larger set of moves and determine the best course of action. As these algorithms begin to collect data over moves and games, they can develop better strategies and use them to beat their human opponents.
Today, many companies have tried their hand in different, modern games. With Google’s AlphaGo and openAI’s Dota2 bot, these AI programs overthrew the game’s best players despite having to wrangle with a lot more variables and combinations of moves. This has been possible only due to developments in hardware and computing capabilities.
Self-driving cars have many underlying technologies that power it but AI plays a predominant role in improving the entire experience. While this isn’t just limited to the Teslas and other self-driving cars that most of the public can’t afford, AI is already behind some of the functions we use today like self-parking, lane-assist, and cruise control.
Despite the abstractness and irreproducibility of art, if fed sufficient data, AI models can replicate almost all art forms that humans thrive in like music, pieces of literature, paintings, acting, and more.
AI is already used for music creation and drawings. With Amper, A.I Duet, and more for music production and tools like LANDR for post-music technicalities (mixing and mastering), AI is becoming an instrumental part of a musician’s life.
Similarly, today’s cutting-edge AI is being used to compose literature sonnets, digital masterpieces, and even entire movie-script. In the future, seeing this integrated other technologies like robotics (for painting) and deep fakes (for movies and plays) could redefine art as we know it.
(if this topic interests you, check out this other post I wrote where I cover the intersection of ART + AI in greater detail.)
Trading algorithms are now being powered by AI. Already, a firm called Nomura Securities is making predictions on current market trends based on compiled trading actions of its past traders. It does so by drawing parallels to what happened in the past, in similar economic situations. This is neither the first company to adopt AI-powered models or even algorithmic-based stock trades.
Already, forms of technology dominate the trading sector with high-frequency trading becoming more and more popular. In the last 10 years, actual humans buying and selling stocks has become extinct. Algorithmic trading is the norm. While algorithmic trading consistently provides marginal gains compared to humans trading (through a thorough analysis of stock price variation with respect to time as seen below), AI trading platforms are the next wave. These will likely be able to replace the functions of a human trader and would also remove the emotional aspects that distort trading.
Image by Sabrina Jiang © Investopedia 2020
As the images suggest, through more thorough investigations of stock prices, machines may find a more optimum time to buy and sell stocks (for a much smaller window of time as well) that simply isn’t natively intuitive for humans to detect. In fact, due to this, the average time a stock is held dropped from an agonizing eight months in 2000 to just under 22 seconds in 2011. However it isn’t all rainbows and unicorns (😉), these forms of trading can malfunction. In fact, in May, 2010, there was a flash crash that lasted just 36 minutes in which time the Dow Jones fell by a whopping 9% (and almost immediately recovering back fully). While these high frequency trading algorithms cannot be completely blamed, they had a significant role to play. Following this, there were an onslaught of regulations and it’s likely that the same will follow once AI-powered trading algorithms become mainstream. I only hope that the necessary oversight is taken before business are unnecessarily affected by a similar lapse.
Some startups and banks have also come up with personal financing solutions that range from tasks as minuscule as reminding you when to pay, which actions could help your credit score or even as broad as portfolio management.
An issue, however, since the topic of finance is sensitive is that of data anonymity. As studies have shown, even if data is protected, attempts can be made to match identities to data sets.
Possibly one of the most frightening uses of AI is in the military. Many governments worldwide are keen on leading this arms race. Even though it is mainly governments that are developing these weapons they desperately need and have asked for collaboration from top AI researchers and Big Tech but both have increasingly expressed their ethical concerns. Critics of AI have been most worried about this use citing that it will be the most dangerous for mankind.
There is sufficient evidence to back this. According to a war game conducted by the Pentagon, an AI combatant was able to withstand 5 human combatants. This in itself seems like an overwhelming jump from current standards not to mention that AI can also be used in unmanned armed drones, flight and missile assist, weapon maintenance, autonomous weapons and so much more.
Seeing that the use of AI will become more prominent in so many domains, there is no point in just fearing it. AI is inevitable and will disrupt almost ever sector it can reach.
With many of these applications our lives will unequivocally be transformed. Better universal healthcare, safer cars, easier access to companionship and more point towards a utopian world with vastly improved standards of living. Other uses cases like gaming, chatbots, space exploration and artificial creativity may seem trivial but convenient. Again contributing to an idealistic future. However some implementations of AI like in finance and military are, to put it lightly, worrisome. While they both may seem lucrative in the short-term as countries fight to prove their economic and defence dominance, there are significant red flags that will arise if its development and use continues unrestricted. The risks that come with it like a loss of personal privacy, identity fraud shouldn’t be brushed aside, but even still pale in comparison to the destructions of races and possibly even humanity. Don’t believe me? Even Stephen Hawkings said so!
I think that some of the most important decisions that humanity will make this century will be around controlling the use of AI. I would rather prefer precautionary measures rather than reactionary ones even if this means getting to that utopian world a little later.