Artificial intelligence is the future, but is it a good one?
Ever since the very idea of artificial intelligence (AI) was coined, opinions on it have been divisive, to say the least. Science fiction authors and futurists alike had and still continue to have field days about the matter, wondering in a veritable multitude of ways as to how machines capable of ‘thought’ would be able to change the world and bring about a future very different from what we are used to – be it one of glory and prosperity, or death and destruction.
Scientist and tech visionary Andrew Ng once quipped, “Artificial intelligence is the new electricity.” And indeed, just like how the advent of electricity transformed society over the course of the last century, artificial intelligence has already started to shape our view of reality in ways that most of us don’t even realize yet. While a truly ‘conscious’ and/or ‘sentient’ AI is yet to emerge, AI programs that are capable of ‘learning’ and improving upon their own processes have already been developed, and they have been put to use in a myriad of technical fields. And this time, it seems that the paradigm shift caused by the introduction of AIs to the world as we know it may have seismic repercussions if it is not carefully regulated.
Artificial intelligences have long been favourite plot points of science fiction movies and fiction, and we are all familiar with fictional big-name Hollywood AIs such as Skynet and HAL 9000, many of which are showed to rapidly reveal a malevolent personality that threatens the world at large, or at least the people around them. While AI systems of present are yet to present a personality that goes beyond crude emulations of human ones, without extensive failsafe mechanisms in place, the danger posed by militaristic use of artificial intelligences can be a very real one, for reasons that can be as simple as data bias, a lack of ‘empathy’, or simply because a machine would not process all the aspects of a threat the way a human would, and is likely to respond with countermeasures that may endanger humanity as a whole. Despite these concerns, many countries have been pouring billions of taxpayer dollars into developing AI-based weapons technology, with USA and China at the forefront. Elon Musk, one of the most prolific technology-focused entrepreneurs of our time, has gone as far as to say that AIs pose an ‘existential risk to humanity’ that simply cannot be ignored.
Even worse is the fact that AIs don’t necessarily need access to weapons systems to wreak havoc. Recently, OpenAI, an AI research firm of which Elon Musk himself was recently a member, unveiled an AI that is capable of spinning volumes of extremely convincing-sounding stories and fake news based on a few pieces of (mis)information. In a world where war using information is as critical (if not more) as war using weapons, the threat posed by such technology cannot be understated. Responsibly, OpenAI withheld the full version of the program from falling into anyone’s hands, but it is safe to say that if they have managed to develop it today, it is only a matter of time before another entity with less benevolent intentions comes up with another system that does the same job, and puts it to use. Many of the biggest names in technology are already dabbling in the field of AI, and it is safe to assume that if any of them develop such a technology, they would not be hesitant to sell it out to the highest bidder.
Speaking of warfare on the informational front, ungoverned and irresponsible AI research can also lead to the development of new computer viruses that would be capable of masking their presence by intelligently avoiding behavioral patterns that antivirus programs are designed to detect. Such programs would even be able to alter their own code to become more effective, allowing it to bypass security measures or to simply cause even more damage. More comprehensive cyber-attack AI platforms would be able to study and learn the behaviors of the users of their target machines, and deduce how and when to carry out attacks of devastating potential. Combating such threats would require the development of new AI-based antivirus and cyber defense solutions, which would not only require a massive drain of resources to be made, but would also require specialized and expensive hardware to be effectively run.
However, what appears to be scariest aspect of the AI revolution in an everyday context is that it is rapidly growing more and more complex, and its rapid adoption by companies would mean that in another decade or so, we would be having real-time conversations with AI machines without even realizing it half the time. Google has already demonstrated a working prototype of such a system. Such AI programs can be godsends to scammers and manipulators, who would be able to use them to their complete advantage. Considering that this is an age where virtually everybody is busy uploading the most intimate details of their lives into complex stratified databases in the form of social media posts, using said information with malicious intent to fool people into doing the biddings of the highest bidder is merely the next logical step. Machine learning technology, as its stands today, has already displayed prototypes of very real-looking and real-sounding news anchors who are in fact digital simulacra that can be customized to suit specific needs of clients.
While there is no concrete indication that artificial intelligences would spell certain doom for humanity, the threats they pose are very real indeed, and without proper governing and safeguarding measures, there is ample room for AI systems to be misused. The fear of what we do not fully know or understand yet is only natural, and in this case, at least until public awareness about the potential dangers of AI becomes more mainstream, that hesitation brought along by the fear may become a positive decisive factor behind the weight of their ultimate impact.
