February 24, 2016
Siri’s Path to World Domination: When Our Minds Become their Minds
Aasimah Tanveer
Department of Biology
Lake Forest College
Lake Forest, IL 60045
Download PDF
Artificial Intelligence (AI) might seem like a science fiction daydream when, in reality, we live and work with AI everyday. Our self-driving cars, shiny medical machines, even Google are all possible because of AI. When casually mentioned, AI immediately brings images of popular cultur-al portrayals such as Hal 9000 in 2001: A Space Odyssey, Cortana from the “Halo” videogames series, R2D2 and C3PO from Star Wars, JARVIS and Ultron from the Marvel Universe, or Robocop himself. The portrayal of AI is future-oriented leading people to think that full implementation is impossible to reach in our lifetimes. However, now, more than ever, we are on the cusp of creating a being even more sophisticated than ourselves.
Although many people associate AI with robots, a robot is just a container for the artificial intelligence technology itself. The AI is the brain of the robot body. John McCarthy, publisher of the first major paper on AI, coined the term in 1956 and observed “as soon as it works, no one calls it AI anymore.” Meaning that people take AI technology for granted once it is conveniently integrated into their daily lives, like the apps on a cellphone. McCarthy defines artificial intelligence as “using computers to understand human intelligence, but AI does not have to confine itself to methods that are biologically observable.” He defines intelligence itself as: “the computational part of the ability to achieve goals in the world,” essentially stating that AI is a goal-oriented, computational process to understand and mimic human intelligence.
Since the concept is so extensive, scientists have defined three main levels of AI: artificial narrow intelligence (ANI), artificial general intelligence (AGI), and artificial superintelligence (ASI). In order to achieve synthetic human beings, we must progress from ANI to AGI and finally, to ASI. Artificial narrow intelligence, also known as “weak AI”, is computer intelligence that specializes in one area or goal. Everything from the navigation system in our cars, to using Google to find the right webpage, to the health tracker on the treadmill, is a form of AI that we interact with on a daily basis. Email spam filters, phone apps, stock trading software, and even the recommendations that pop up while surfing the internet are all products of an ANI algorithm geared towards making our lives easier, one task at a time. The smartest ANI computer currently is IBM’s Watson computer, the famous computer constructed to win Jeopardy! Finding sexy singles in your area is the first step to a supercomputer.
The next milestone to achieve in AI is to build a brain, or artificial general intelligence (AGI). This “strong AI” is achieved when the machine is just as intelligent as the average human in all tasks. A computer, at this level, should be able to perform all the mental problem solving, decision-making, and general heuristics of the average human being. We, as the most highly developed beings on this planet, take for granted the fact that we do not have to sift through entire, inefficient algorithms to make a decision. In addition to efficient algorithms, we use mental heuristics, or short cuts, that allow us to make split second decisions based on underlying cues and gut feelings. A computer must go through a purely logical process of parsing through data with specific, defined parameters, which takes much longer than the human brain. Humans also have a highly advanced learning system called neuroplasticity, which allows our brain to become faster and pick up tasks more efficiently the more it is used. Machine learning is currently not advanced enough to expand on itself in this way. A programmer must specifically give the code for the machine to use previous history of use to modify its parameters.
Scientists around the world are working on achieving AGI. The Obama administration has allocated over 300 million dollars to the BRAIN initiative, which is working on reverse engineering the human brain while Europe has its own Human Brain Project with a billion dollar budget to achieve the same “build a brain” goal. Moore’s law states that processor power doubles every eighteen months so more time and resources are needed to create more refined, genetic simulating algorithms than are needed to even come close to resembling a structure that has been processed through thousands of years of evolution.
The most juicy and controversial topic of AI is artificial superinteligence (ASI) as the possibility of full integration becomes near. Nick Bostrom defines superintelligence as ” an intellect that is much smarter than the best human brains in practically every field, including scientific creativity, general wisdom and social skills.” Ray Kurzweil, a noted and controversial computer scientist, refers to this stage as “singularity” or “transcendence” The culmination of ASI is when computers can replicate and evolve themselves by writing their own programs using their superior intellect. In order to achieve ASI, significant hardware and software improvements from AGI technology must be achieved such as increased speed, size, storage, reliability, and durability on the hardware end; and the ability to self edit and self upgrade on the software end. Innovations in the very structure and execution of software to create collective capabilities and quality intelligence are needed as opposed to solely increasing capacity. Scientists have still not been able to understand the soul or find a way to empirically determine its existence. In order to create a human being, such factors must be known and primitive yet complex concepts, such as emotions, must be programmable.
These may seem like too many tasks to achieve and naysayers claim that all these obstacles make it impossible to achieve ASI in our lifetime; however, the scientific community has made more advances in the last fifty years than we have in all preceding human history combined. Humans have the tendency to look to the future linearly and extrapolate the future based on evaluating past experiences, but technology, as seen with Moore’s Law, progresses exponentially. From this point, it seems like a tall order, but we are just a point on the graph without full perspective of the steep upward curve ahead. This phenomenon, named the law of accelerating returns by Kurzweil, is what makes ASI entirely possible to come to fruition in our lifetime. Once ASI is achieved, the breakthrough will change the very fabric of our society.
The future applications of developed AI algorithms are endless. Michio Kaku mentions the possibility of having in-home robotic health professionals that closely monitor and care for the sick, like a real-life Baymax. Andre Leblanc posits that researchers can test new drugs on simulated, virtual human beings created by advanced AI algorithms in order to eliminate the risk associated with clinical trials and to retrieve quicker results. Virtual reality gaming is emerging as a new method for immersive gameplay with products such as the Oculus Rift, but with more advanced AI technology, improvements can be made to advance beyond even advanced augmented reality gaming. A whole new market of immersive entertainment has the potential to flourish with Matrix-like worlds to capture consumers. ASI technology advancement will see the rise of mental data sharing, downloadable experiences, and artificial companions. Jobs will continue to be replaced by these “smart” machines while humans are free to explore other areas of study and leisure. Achieving artificial superintelligence could be the pinnacle of human existence, or the inevitable downfall.
Ethical questions also arise if such sentient beings can be synthesized. If machines can be human, what distinguishes us from them? Are they equal in status to us? Will they be the white-collar workers or great minds working on breakthroughs with their superior intelligence? Integration of a completely foreign concept usually does not happen smoothly in society, so it might result in a District 9 situation, or we might even be ruled by these hyper intelligent beings. Even the ASI technologies themselves would change the economy with groundbreaking innovations in both consumer and in- dustrial markets. Experts on both sides of the “should we advance further into AI” topic speculate immense possibilities, both beneficial and dreadful.
Artificial intelligence, as frightening and magnificent the concept, is revolutionizing the way we conduct our lives. Immense efforts are being allocated to the field, which is the place to be for scientists in this technology age. You could love a robot, program your child, or visit a new galaxy in your very head. Even science fiction’s imagination falls short of what the future will be.