The Foundations and Future of Artificial Intelligence
The Reality of AI
Artificial intelligence tends to conjure fantastical images of sentient mechanical beings from movies such as Chappie or iRobot, but the reality of AI today is more understated yet far-reaching than Hollywood might suggest. In fact, you already interact with artificial intelligence on a day-to-day basis and may not even realize it. Common experiences such as emailing, simple web searches, or getting directions are all built on advanced AI algorithms.
Think back on all the times Netflix, Amazon, or YouTube has recommended a show or product to you. All of those recommendations are built on formulas that examine what you buy or watch and learn to offer suggestions for new products you might enjoy. The driving force behind these algorithms and recommendations is AI. It’s not hard to imagine why this technology is so critically important for many businesses. In fact, AI has the potential to contribute upwards of 15.7 trillion dollars to the global economy and increase GDP output for local economies up to 26% by the year 2030.
Building the Brain
In 2011 Google began developing an AI platform capable of deep learning aptly named Google Brain. The Google Brain is a network of both hardware and software that mimics the process of real neurons to automatically learn and improve based on experience. This network analyzes vast amounts of data and learns how to perform all kinds of handy tasks such as recognizing and responding to commands spoken into a smartphone, identifying objects and places in photos, and responding dynamically to internet search queries. As it turns out, this type of learning neural network is capable of performing many tasks faster, more efficiently, and on a greater scale than humans can.
Even simple things such as filtering out spam emails are based on these dynamic machine learning systems. The efficiency at which these algorithms function is substantially greater than manually updated software rules because it is capable of learning what constitutes spam directly from the content of the email. This sort of machine learning also gives users their own unique and personalized experience as it builds a model based on your preferences. What is spam to one person may not be spam to another, and the Google Brain is capable of situationally differentiating between the two.
The Brain Behind Your Virtual Assistant
How often do you ask Siri a question? Or update your calendar with Alexa? Or schedule an appointment with Google Assistant or Cortana? Any of these sorts of interactions are stored in the cloud so it may then learn from and improve future results. In 2015 Apple stated that Siri was processing an average of one billion requests every single week. That’s just shy of 100,000 interactions every minute and this is only one of the several large AI platforms. It is quite likely that number has increased substantially since 2015, and this means that these sorts of deep learning AI algorithms can tap into unfathomable amounts of data to learn from. The more data that accumulates on any given person, the easier it is for the AI to recognize patterns in behavior and consumption which allows it to better predict the end user’s needs.
Artificial Intelligence and Bank Security
Besides the obvious potential for AI to take over the more mundane clerical side of banking, its ability to detect fraudulent activity by instantly recognizing anomalous behavior is extremely attractive to any financial institution. Banks can use AI to monitor transactions and spot suspicious activity via intricate pattern analysis at a high volume with surprisingly accurate results. According to a 2019 Business Insider report, the aggregate potential cost savings for banks from artificial intelligence applications is estimated to reach nearly $450 billion by 2023. This implementation of AI serves not only as a massive cost saver for the banks themselves, but also helps protect consumers from losing their money to potential scammers.
Artificial Intelligence Drives Progress
Self-driving automobiles are already making an appearance on the road and given enough time chances are high that they will become the standard. As these cars drive around and interact with other vehicles (both self-driving and otherwise) they continuously utilize machine learning to improve. Given long enough these sorts of cars will be able to analyze situations in real-time; they will be able to take what they’ve learned in the past and apply it to current conditions in order to make better decisions. While this sort of ‘autopilot’ system is a rather new addition to automobiles, it is already a staple of the aviation industry. If you’ve ever flown on a commercial airline, you indirectly experienced artificial intelligence as it does the vast majority of the piloting. In fact, AI is such an intrinsic part of flying that Boeing 777 pilots spend on average a mere 7 minutes manually piloting the plane. Airbus pilots spend half that. Considering the FAA handles more than 16 million flights every year, it is safe to say that AI has had a widespread yet subtle impact on the daily lives of most Americans.
The (Realistic) Future of Artificial Intelligence
These ‘basic’ forms of AI are laying the foundation for more advanced iterations of machine learning in the future. Barring some massive unforseen technological shift, I sincerely doubt we will be seeing any walking, talking, feeling, robots in the near future. However, the combination of AI with new and burgeoning iOT technology creates some really interesting possibilities. On the more mundane side, imagine if Siri could tell you where the nearest available parking space is, or if your smart-fridge could detect when stocks are running low and reorder groceries automatically for you. On the more futuristic-yet-possible side, projects like Neuralink from tech mogul Elon Musk seek to physically integrate AI into the human body.
Neuralink is an incredibly ambitious project that wants to physically implant ‘brain-machine interfaces” (BMIs) into the brain of the user in order to enhance their life. Initial steps are targeting the treatment of strokes and serious brain diseases, giving people with impairments the ability to perform basic tasks such as controlling a robotic arm, typing on a computer, or changing a television channel using only their thoughts. Ultimately, Elon is chasing human symbiosis with artificial intelligence, with the eventual goal being complete human enhancement.
Currently, some neuroprosthetics have the ability to interpret brain signals that allow basic control over their prosthetic limb. Elon seeks to take this technology and expand upon it by linking these implants to external devices and software via high bandwidth broadband internet connections. The implications of this type of technology are profound, and are viewed by some (including Musk) as a way for humans to compete with the advancement of AI without becoming irrelevant in comparison. Envision being able to download information from the internet directly to your brain, allowing you to learn new skills or access information instantly. Can you imagine setting a download for a new language before going to bed and waking up being able to speak it fluently? It’s an image that harkens back to a scene from the Matrix where Neo downloads Kung Fu directly into his brain. A mere twenty years ago this technology was featured by Hollywood on the big screen as something so futuristic and out of reach, and yet now it is being developed by private sector companies with the goal of putting it in the hands (or brains) of the average person in the very near future.
At Lithios we value outside opinions. This blog was written by one of our guest bloggers.