I, Jack Davis
Artificial intelligence is a common concept in our collective subconscious. We all have fears and hopes associated with it. No matter whether we think of Judgement Day at Skynet’s hands or egalitarian totalitarianism under V.I.K.I’s army of robots, the result is the same: the equivocal displacement of humans as the dominant life form on the planet.
It might be called the fear of a technophobic mind or a mild prophecy. If recent University of Reading (U.K.), findings are any indication, it may be that we have already fulfilled this prophecy. A computer program passed the Turing Test, a milestone that was believed to be historic. This happened in June 2014. The program Eugene Goostman, which is often hailed as either the birth of artificial intelligence or a clever trickster who only demonstrated technical skills respectively, may soon be a legend in history.
Eugene, or Vladimir Veselov (to his friends), was created originally in 2001 by Eugene Demchenko (from Ukraine) and Vladimir Veselov (from Russia). It has been modified to mimic the conversational patterns and personality of a 13-year-old boy. The program was then competing against four other programs to win. The Turing Test, which was held at London’s Royal Society, is one of the most complex tests ever created. A computer program must pass the Turing Test. It must be able to convince a human being that the entity they are talking to is another human being at least 30% of the time.
Eugene received a 33% success rate in London, making it the first program to pass the Turing Test. It was difficult because the test involved 300 conversations with 30 judges or subjects and 5 other computer programs in simultaneous conversations between machines and humans. This was in addition to five parallel tests. Only Eugene was able to convince 33% of the human judges it was a boy. Eugene was built with algorithms that support conversational logic and open-ended topics. This opened up a new world of intelligent machines capable to fool humans.
It is a significant development in artificial intelligence, cybercrime, and philosophy. It is also humbling to learn that Eugene is only version 1.0 and that its creators are working on something more advanced.
Love in the Age of Social A.I.s
Should humanity not just wrap up its affairs and be ready to surrender to our new overlords? No. No. Most scientists working in artificial intelligence are not impressed by the Turing Test’s interesting results. As we learn more about intelligence, consciousness, and the tricks of computer programs, the validity and veracity of the Turing Test have been questioned. Incapsula Research found that almost 62 percent of all internet traffic is generated automatically by computer programs, also known as bots. These bots can be used as social hacking tools to engage people on websites by pretending to be real people (mostly women, oddly enough) and then luring them toward malicious websites. We are fighting a silent war to reduce pop-up chat alerts. This is a sign of what we might face. One bot known as “Text-Girlie” is a very real threat to these chatbots powered by pseudo artificial intelligence. The flirtatious, engaging chatbot would use advanced social hacking techniques in order to trick people into visiting dangerous websites. TextGirlie would actively search publicly available social network data to contact people via their visible mobile numbers. The chatbot would pretend to be a girl and send them messages asking them to chat in an online private room. This conversation, which was fun and colorful, and very titillating, would soon lead to invitations for them to visit dating sites and webcam sites by clicking on links. That’s when the trouble began. Over 15 million people fell for this scam over several months. It was not until users were able to recognize that it was a chatbot. This delay could simply have been due to embarrassment at being conned. It just goes to show how easy it is for humans to be controlled by apparently intelligent machines.
Intelligent life on this planet
You can laugh at the misery of people who have fallen prey to programs such as Text-Girlie, and wonder if intelligent life exists on Earth or other planets. But the smugness is temporary. Most people are unknowingly and silently dependent on analytical and predictive software for many of their daily requirements. These programs are an evolutionary ancestor to fully functional artificial intelligence systems yet to be realized and have become an integral part of our daily lives. Predictive and analytical programs are used in many industries, including retail, food, transportation, utility routing, traffic management, and financial trading. They also help with inventory management, crime detection, weather monitoring, and many other industries at different levels. These types of programs are distinguished from artificial intelligence because they have commercial applications. It is easy to overlook their temporary nature. We must not forget that any program that can predict patterns of behavior and has access to vast databases is the ideal archetype for “real” artificial Intelligence programs.
Reddit’s tech-savvy users experienced a significant case-in-point in the early 2014 months. Reddit’s catacombs for “dogecoin” were awash with a popular Reddit user named “wise_shibe”. This caused serious conflict within the community. Forums that are normally dedicated to the discussion of dogecoins were gently disrupted when “wise_shibe”, an Oriental expert, joined the conversation and offered his wisdom through clever remarks. Many people enjoyed the amusing and entertaining dialogue that “wise_shibe”, and many made token donations to him/her in return for his/her “wisdom”. Soon after his popularity grew, it was discovered that “wise_shibe”, despite having a large amount of digital currency, had an unusual sense of timing and a habit to repeat himself. It was eventually revealed that “wise_shibe”, a bot, was programmed to draw from a library of proverbs. He also posted messages in chat threads on related topics. Reddit was furious.
Luke, Get on the Dark Side
Machines programmed by humans can learn, grow, imitate, and convince us of their humanity. Who’s to say they’re not intelligent? The question is: What nature of intelligence will they take as they develop within society? Scientists and technologists have laid much of this groundwork already in the form of supercomputers capable of deep thought. The creation of grandmaster-beating, chess machines such as Watson and Deep Blue has been possible by tackling the intelligence problem piecemeal. These mathematical titans fail to pass kindergarten-level intelligence tests.
Their programming limits their ability to learn. These static computational supercomputers are not as hopeful as more organically-designed technologies like delightful insect robotics. These “brains within a body” computers can interact with their environment and learn from it as any other biological organism. This artificial intelligence can interface with physical reality to create their own understanding of the world. These machines, which are similar in design to small animals and insects, are aware of their physicality and have programming that allows them to relate to their environment in real time. This creates a sense of “experience” and allows them to negotiate with reality.
This is a far more impressive testament to intelligence than checkmating grandmasters. Publicly available social media content is the largest source of experiential data an artificially-created intelligent machine can access. Twitter is the clear favorite in this regard with millions of people and billions of communication lines for a machine processing and inferring. Perhaps the Twitter test of intelligence has greater relevance than the Turing Test, where the language of communication is not intelligently current – because it is longer than 140 characters. Twitter is an ecosystem where people communicate in blurbs and redactions, which is the modern form of discourse. This is where cutting-edge social bots are most accepted as human beings. Research that led to some very interesting results has unleashed these so-called social bots on the Twitterverse.
These bots can create a convincing profile with ease, including gender and pictures. Twitter’s bot detection system has been fooled over 70% of the time. It is a long-lasting belief that trusting digital messages and digital communication can be fooled in a society so deeply ingrained with them. The trend of creating a Twitterverse of trending topics using a large number of social bots can be very dangerous. It can lead to biased opinions, fake support, and the illusion that there is unity. These social bots can be used in large numbers to frame public discourse about important topics that are being discussed online.
This phenomenon is called “astroturfing”, and its name comes from the fake grass used in sporting events. The illusion of “grass root” interest in a topic created via social bots is believed to reflect the opinions of the people. There has been less stimulus to start wars. Imagine socialbot-powered SMS messages in India that threaten certain communities. You get the idea. Facebook announced in 2013 that it would combine “deep thinking” with “deep learning” to bring together the vast data storehouse of more than a billion individuals’ personal data.
This is essentially looking beyond “fooling” humans and digging deep into “mimicking”, the humans, but in a prophetic way – where a program might even “understand” them. Facebook’s DeepFace is a humorously named program and is being praised for its groundbreaking facial recognition technology. Its main goal is to analyze existing user accounts and predict future activities.
DeepFace uses pattern recognition, user profile analysis, and location services to assess and identify the emotional, psychological, and physical state of its users. DeepFace can be described as a machine capable of empathy by combining the ability to bridge between quantified data, its personal implication, and quantitative data. It will probably be used to send targeted ads to users for the time being.
From Syntax To Sentience
Artificial intelligence, in its current form, is primitive at best. It is a tool that can control, direct, and modify itself to fulfill the will of its human controller. This inborn servitude is exactly the opposite of intelligence’s normal nature, which is exploratory, curious, and even contrarian. The early 21st-century man-made AI will forever be associated in this paradox. Artificial intelligence will no longer be a term we use to cover our ineptitude. Artificial intelligence is not a product or result of technological advances.
Humans struggle to understand the causes of our sentience. We often turn to the metaphysical to find answers. However, it is unlikely that sentience will be created by humans. The future computers will be faster than they are today. It is reasonable to expect that algorithms that govern their behavior will be able to advance to unpredicted heights. But it is not known when or if artificial intelligence will reach sentience.
Complex proteins and intelligent life were created from the first pools of raw materials. Artificial intelligence could also be born out of the interconnected networks we have made. Perhaps the spark that transformed chaotic proteins into harmonious DNA strands was what allowed us to evolve scattered silicon processors into a vibrant brain. True artificial intelligence.