Is there a possibility of AI being backed by morality?
Recent advancements in artificial intelligence have demonstrated that our computers require a moral code. At first look, the aim is easy enough—create an AI that behaves morally responsible; yet, it is significantly more hard than it appears, since there are an incredible number of elements at play.
While the public may be concerned about ensuring that rogue AI do not decide to wipe out mankind, such a danger is not a current concern (and will not be for some time). So, how do you create an AI capable of making complex moral decisions?
Identifying the Issue Of Combining Artificial Intelligence & Morality
There’s a two-pronged approach to this. Making ethical decisions in order to identify patterns and then working out how to put it into artificial intelligence. On the most basic level, it is all about predicting how a human would behave in a particular situation.
However, one important issue is that our moral judgements are not objective—they are neither eternal nor universal. Maybe our moral evolution hasn’t reached its zenith yet, and people a hundred years from now would think that part of what we’re doing now, like the way we treat animals for instance, is absolutely wrong.
As a result, there is a risk of prejudice and becoming trapped at our current level of moral growth. And, obviously, there is the previously acknowledged issue of moral complexity.
Bias exists in facial recognition systems and digital assistants against women and people of color. Despite widespread use of artificial intelligence, social media platforms such as Facebook and Twitter struggle to restrict hate speech. Algorithms employed by courts, parole offices, and police agencies create arbitrary parole and punishment decisions.
A growing number of computer scientists and ethicists are addressing these concerns. It is critical to use methodologies from computer science, philosophy, economics, and psychology to tackle these challenges and figure out exactly how morality operates and can (ideally) be written into an AI.
AI Like Humans
We can design systems to accomplish many things that humans do well, but other tasks are extremely complicated and difficult to convert into a pattern that computers can identify and learn from.
The fact that many things we believed were meaningful standards for intelligence, such as being able to play chess successfully, were really relatively accessible to computers throughout the first two decades of AI study.
Writing and developing a chess-playing program was not straightforward, but it was achievable. Indeed, today’s computers can defeat the world’s greatest players in a variety of games, including chess and Go.
Consider the following: When a youngster on a bicycle swerves in front of a vehicle traveling down the road, the car is forced to stop. Does the car drift into the oncoming lane, colliding with another vehicle? Does the car go off the road and collide with a tree? Does it keep going and hit the youngster?
Each solution has a drawback: It might lead to death.
It’s a terrible circumstance, but humans encounter such situations all the time, and if an autonomous automobile is in charge, it must be able to make this decision. That implies we’ll have to figure out how to implant morals into machines.
Conclusion
Many of the decisions they make have an influence on individuals, and we may need to make judgments that are regarded ethically weighted. But, of course, knowing what option to make requires first understanding how our morality works (or having a solid concept). We can then start programming it.
The ethical difficulties raised by AI in surveillance extend beyond the simple collection of data and the focus of attention to the use of information to control behavior, both online and offline, in a way that undermines autonomous rational decision.
Of course, attempts to control behavior are not new, but they may take on a new dimension when AI systems are used. Users are prone to “nudges,” manipulation, and deceit due to their intense engagement with data systems and the rich information about persons that this gives.
