Artificial Intelligence: The Ethical Conundrum

We’re ethically unprepared for the rapid advances in AI and machine learning and their unexpected consequences.

Machine learning to-date

We have come a long way since the first computer was introduced. In fact, the first prototypes of computers had smaller computational power than our modern smartphones and were as useful, albeit much bulkier, as a fancy calculator. In today’s world, machines have gotten smaller, faster, less expensive, and… smarter. They are no longer learning in one-dimensional ways, by following exact algorithms, but instead do it by churning through vast amounts of data and assessing the likelihood of certain combinations and probabilities, rather than offering unilateral solutions to complex scenarios.

Computers outsmarting humans

This enhanced machine intelligence is certainly opening up a new world of possibilities for mankind and enabling us to advance in fields like manufacturing, aeronautics, science, biotech or medicine at much faster rates. We are able to process amounts of data previously unthinkable and thus, we can also uncover patterns and trends that we simply had no chance of discovering, equipped only with the limited human brain processing capacity. In a way, machines are starting to work in multi-dimensional patterns similar to those of brains, which in itself isn’t alarming, lest it resulted in unexpected and unwanted outcomes.

The threat of AI

Ever since machines got smarter and more sophisticated at recommending solutions, people got more concerned about them taking over our lives, jobs, cars, homes, even our loved ones. Movies like I, Robot, Ex Machina, Lucy, Her, Ghost in the Shell, would have us believe we are on the verge of an inevitable existential crisis, where all human-run functions in our lives are to be replaced by operating systems, voice recognition assistants, artificial intelligence, and our realities – converted into a seamless blend of virtual and physical spaces, with more emphasis on the former. 

What’s ethics got to do with it

Corporations and large businesses are already mining and using big data from various sources to make subjective decisions in all work areas: HR, procurement, financial planning, marketing, logistics, and more. As AI becomes more sophisticated and harder for the human brain to grasp, it also becomes increasingly harder to control. Machine error is different from human error and cannot always be detected, let alone predicted. Over-reliance on this type of human-free mechanisms can eventually lead to catastrophic outcomes of various proportions, especially in fields where human lives, health, safety or education are concerned. As techno-sociologist Zeynep Tufekci exclaimed during her remarkable Ted talk, "We cannot outsource our responsibilities to machines. We must hold on ever tighter to human values and human ethics."

Machine intelligence era

Modern computational systems can infer virtually anything about anyone from our digital footprint, even if we haven’t willingly disclosed it to any organisation, just based on our public social media data, for instance. Hidden biases in black box algorithms discovered by researchers have impact and real-life consequences that can’t be ignored as we become more reliant on machine-run systems. Making super intelligent AI is a huge challenge and making it safe is an even more complex process. These concerns should be anything but an afterthought when empowering machines to be smarter, faster and more helpful.

Rumour has it, Donald J. Trump was able to cleverly target swing voters during the recent U.S. Presidential election by analysing their social media data, and serving them appropriate pro-Trump and anti-Hillary advertising. Fair? That’s certainly up for debate, however, it could be argued that the party/business/individual with the best big data analysts can benefit from a significant advantage that is largely based on good marketing and not much else.

Human vs. machine

In the end, decisions are only good as the data behind them, and no computer-run process, be it artificial intelligence or not, is smart enough to be able to discern things like gender bias, creativity, innovation, or other qualities that require EQ (emotional intelligence) – a realm that is (still) best left to us, humans.

Copywriter: Ina Danova
All articles

We develop Enterprise Apps and IoT solutions that bring together devices, cloud & data, sensors, etc. for your business needs. We develop software (mobile apps, api’s, etc.) and hardware to connect devices and machinery in various kinds of industries.

Get in touch