But the reality is much more nuanced. As artificial intelligence becomes more integrated into our daily lives, it is increasingly important to think about the rewards versus the risks of employing AI architecture.
Safety First But Not Foremost
As we explore more applications of AI, we will reveal more dangers that must be addressed. Like space travel, risks exist but we must persist in order to learn and improve. In any pioneering enterprise, the risks range from tragic to mere flaws. We must tolerate risk in order to gain rewards.
Still, we will not be able to anticipate all the dangers and prevent accidents with the use of AI. As we advance, we’ll need to find levels of acceptable risk and protect against the tragic.
Physical Safety
Advancements in self-driving cars make this technology imminent, but in 2018, one collided with a pedestrian when it could not differentiate the person from other objects. The algorithm has since been adjusted, but it’s hard to justify loss of an innocent life with scientific advancement.
However, the loss was not in vain. Better algorithms developed. The ability to detect objects on the road improved technology like brake-assist, lane-change warnings, and other methods, keeping humans from making mistakes that lead to collisions. Many lives have been saved.
Technological Safety
AI poses an equal level of risk and reward when it comes to data. The question of safety is a double-edged sword: social media, businesses, credit reporting services, and schools are all using more complex programs to maintain billions of tidbits of information about us as individuals, families, social groups, and nations, but malicious AI can hack algorithms and compromise billions with the click of a button
Further muddying the waters is that AI algorithms are also used to stop the hackers. Hackers are more successful with their programs, but safeguards are stronger and more robust for handling attempts.
Do we sacrifice the power that comes from harnessing big data because of criminals? By contrast, society in the Wild West didn’t stop mail delivery or the railroad industry because of bandits…
Societal Safety
Now, what about that possibility for sentient AI to target humans? Movies show machines rebelling against humans once they reach true intelligence; sometimes just by achieving sentience, others by a virus, bad code, or because they feel subjugated.
As advances in AI bring us closer to the eventuality of machines thinking for themselves, will these science fiction plots remain science fiction? There is still time to prepare and build safeguards against the dangers.
What Is Our Obligation To Safety?
The scary truth is that safety is an illusion and that tragedy can strike at any time. However, as a species and society, we work hard to prevent this every day. Laws, choices, and morals guide us away from danger. As we enter into a new era that includes AI, we’ll be confronting new challenges and facing new forms of risk, but we’ll also be discovering new methods to overcoming and handling them all.
