The brain of a human being is far beyond what any machine is currently capable of. While it’s true that a computer can perform millions of calculation much faster than the human brain, humans can perform much more complex activities like showing emotions, balancing their bodies in highly difficult terrains, and adapting to changing environments and times. Artificial Intelligence (AI) tries to bridge the gap between what the human brain and a computer can do.
One of the most successful outcomes of this, is the humanoid robot — Sophia , created by Hanson Robotics. The company goes on to say,
Hanson Robotics’ most advanced human-like robot, Sophia, personifies our dreams for the future of AI. As a unique combination of science, engineering, and artistry, Sophia is simultaneously a human-crafted science fiction character depicting the future of AI and robotics, and a platform for advanced robotics and AI research.
While it may seem a fictional character and a cool concept, it may not be unreasonable to think that Sophia might even behave like you, if given a chance for a long term interaction with you. Have you ever wondered, what if you have a doppelganger AI who is just like you? It may sound both funny and spooky, but AI research may surely enable it in the near future.
AI is a very difficult topic, but it’s looking promising. By the first half of 20th century, science fiction movies brought in the concepts of intelligent human-like robots and computers. However, enthusiasts didn’t take this just as a work of fiction. By early 1950s, scientists and researchers started taking interest in AI . One such person was Alan Turing, a highly influential individual in the development of theoretical computer science and the famous — ‘Turing machine’. In his 1950 paper, “COMPUTING MACHINERY AND INTELLIGENCE”, Alan Turing asked a question —
Can machines think ?
Turing was curious to discover, that if humans can assimilate information from their surrounding to learn and make decisions, why can’t computers do the same ? Since then, AI research has interested a huge numbers of researchers and received enormous funding across the world. According to International Data Corporation (IDC), worldwide spending on AI was $37.5 billion in 2019, and is expected to reach $97.9 billion by 2023 .
Today, one of the major implementations of AI, which interacts with humans is through smartphones, laptops, and digital assistants like Siri and Cortana. They not only help you get done with your daily tasks and set reminders, but also learn your habits, your likes and dislikes, and can even detect your mood from your voice and phone usage. In a 2018 Google IO event, Google demonstrated how their assistant was smart enough to make an actual phone call to a saloon, interact with a human on the other side of the call, and schedule an appointment without a single hiccup.
Though artificial intelligence directly does not comprise of only neural network methods, however neural network and deep learning have become increasingly crucial for the effectiveness of AI . With the advancements in silicon technology, processors have become extremely power efficient, fast and reliable. Moreover, storage technology has evolved a lot over the last decade, where it has become possible to store terabytes of data in blazingly fast pocket-sized SSDs, which was impossible just a couple of years ago.
This has enabled researchers to build and train large complex neural networks  required for AI to emulate human like features such as feelings, and the important act of “reasonable decision making” . Also, a huge number of computations can be made on these, not so power hungry processors, which are enabling better battery life for such remote applications. A huge amount of research is also going on in faster and more accurate sensors and their reliable integrations with onboard embedded systems , which will enable such AI to be more aware of its surroundings than ever before — may be even more than a living human !
Faster and larger storage in a smaller form factor means that a huge amount of information can be stored within the AI hardware. This can be information which was stored by the designers of the AI, but more importantly, it will enable the AI to store information in real time — information that it experiences on interaction with the environment. This information can be used by the AI in real-time to learn new things — what is called “self aware AI” , that is, it knows what to learn and what not. This will enable such AI to adapt to changes and interact more realistically with the environment.
Worth mentioning the emerging 5G tech
5G is beginning to emerge as a super fast network, not only for data consumption, but also near real-time remote control — thanks to its low latency. This means a lot of the AI decision can be made in the cloud, without the need for local storage. Storing and computing decisions in the cloud will not only help in a more apt decision by the AI due to the “almost infinite” knowledge available , but will also help keep track and control of the AI to not perform illegitimate or objectionable tasks. It is worth taking a look at the potential use cases of remote controlling devices using 5G, which was released by Ericsson.
But how a doppelganger ?
How can the AI think and make decisions like me ?
Have you ever talked to your friend about buying an iPhone, and the next moment you see a Google Ad about the latest iPhone? Or were you ever looking to order food, and your favorite dish came up automatically? Or even LinkedIn suggest you a complete message with which you can reply to a connection? — you surely have experienced this !
Huge tech companies like Google and Facebook are continuously collecting your data — what you search, who you talk to, whose posts you like, and even what music your like. This data is used to personalize your experience with their products, to make their products think like you ! But the scope of this information goes much beyond just content personalization.
This data can be specifically used to train AI to behave and decide like you. With the paradigm of connected devices (smartphones and wearables), data about your sleep pattern till people who matter to you, everything in stored in the cloud. A paper in IEEE  also studied the use of Microsoft Kinect, a sensor which detects your posture, for robotic applications. The paper investigated how a robot can be taught indoor navigation by reverse engineering, using data collected from human subjects.
Such AI, if implemented in a robot, can definitely deceive your colleagues for you in an office meeting. It will not only look and behave like you, but in some aspects may also be smarter than you in decision making — thanks to the emergence of 5G and the connected devices paradigm, which will enable the AI to stay up to date with the latest happenings. No, I am not saying that our technology is ready enough for such a move, but we may be close. The rate at which AI is developing and attracting researchers, it is not hard to imagine a future with AI all around us. But, this would lead to the next big question, which has been one of the biggest debates of the century !
All this might seem very cool and helpful to humans for performing tasks and in decision making. But, researchers and the non-scientific community also worry about the potential dangers it may have on us.
- Decision making — Many fear that such smart decision making will make humans too dependent on AI tools to make decisions for them. As this effect will deepen, people will be forced to believe the AI without being aware of the context as to how it works, and will also handicap their independent decision making ability. 
- Cyber threats — Cyber security has been a major concern in the modern era. As people and devices are getting more connected, the vulnerability to cyber-attackers and the potential impacts it may have, is already a startling question. With the sudden influx of AI and its exponential growth, the impacts of cyber-attacks on such technology is not fully known, and the consequences it may have on the society should not be underestimated. Issues such as data privacy also are a major concern, as more and more personal information is used by AI.
- Dangers of autonomous AI — As AI is becoming more complex and smarter, it is believed that soon it will require no human input (or what is called supervised learning) to perform a task. It will be programmed to automatically learn from the environment and decide its action to perform a particular task. However, many fear that such autonomous behavior requires a strict ethical norm to be considered in the AI design , as it should not cause any harm or devastation in completing its objective. A write up by Stanford University beautifully illustrates this potential issue of AI.
AI is either the best or the worst thing to happen to humanity. ~ Stephen Hawking
I would like to conclude this article by asking a question — What would you name your doppelganger? Do let me know in the comments section. I hope you liked this article and it gave you insights on how AI is becoming smart and more aware about its surrounding, and how the data explosion can lead to personalization to such a level, that your office colleagues may mistake your doppelganger for you ! For discussing this more, feel free to email me at [email protected] or DM me on Instagram.