Intelligent machine systems are improving our lives in a variety of ways, including logistics optimization, fraud detection, artistic composition, research, and translation services. Our world gets more productive and hence richer as these systems improve in capability.
Technology behemoths like Alphabet, Amazon, Facebook, IBM, Microsoft, and even famous figures like Stephen Hawking and Elon Musk believe that the time is perfect to discuss the practically limitless field of artificial intelligence. This is in many ways an equally novel territory for ethics and risk assessment as it is for cutting-edge technology. Which debates and issues, then, keep AI professionals up at night?
1. Unemployment. What transpires after jobs are terminated?
The hierarchy of labor’s primary discussion point is automation. As we move from the physical labor that dominated the pre-industrial world to the cognitive labor that characterizes strategic and administrative employment in our globalized society, we have given humans the opportunity to take on more complex tasks as we develop techniques for automating work.
Take trucking as an example; it currently employs millions of people in the United States alone. What would happen to them if the self-driving trucks Elon Musk, the CEO of Tesla, predicted in the upcoming decade become generally accessible? Self-driving trucks, however, appear to be a moral decision if we take into account the lower probability of accidents. The bulk of the workforce in wealthy nations, including office workers, could experience the same situation.
At this point, the issue of how we will spend our time arises. The majority of people continue to rely on selling their time in order to earn enough money to support themselves and their families. We can only hope that by providing this option, individuals would be able to find purpose in non-work activities like taking care of their families, getting involved in their communities, and discovering new ways to contribute to human civilization.
In the future, if we are successful in the transition, we might look back and find it barbarous that people had to sell the majority of their waking hours in order to survive.
2. Inequality. How should we share the wealth produced by machines?
The foundation of our economic system is rewards for economic contribution, which are frequently measured in terms of hourly wages. When it comes to goods and services, the vast majority of businesses continue to rely on hourly labor. But by utilizing artificial intelligence, a business can significantly reduce its reliance on its human staff, which implies that fewer people will receive the profits. Therefore, the only people who will profit from AI-driven businesses are those who own them.
A growing wealth disparity is already evident, with start-up founders taking home a sizable amount of the economic surplus they generate. The three largest businesses in Silicon Valley and Detroit each produced approximately the same amount of revenue in 2014, despite Silicon Valley having 10 times fewer workers.
How do we set up a just post-labor economy if we’re genuinely imagining a post-work society?
3. Humanity. What impact do machines have on our interactions and behavior?
Artificially intelligent bots are getting better and better at simulating social interactions and human discourse. The Turing Challenge was first won by a bot in 2015 by the name of Eugene Goostman. In this test, human raters used text input to communicate with an unidentified party, at which point they had to determine whether they had been speaking with a person or a machine. Moreover, half of the human raters fell for Eugene Goostman’s deception that they were conversing with real people.
The era in which we routinely deal with machines as though they are humans, whether in customer service or sales, is just getting started with this milestone. Artificial bots can devote practically unlimited resources to fostering relationships, in contrast to people who have limits on the amount of time and consideration they can devote to another person.
The reward centers in the human brain can already be activated by computers, even though few of us are aware of this. You only need to look at video games and click-bait headlines. A/B testing, a basic algorithmic technique for content optimization, is frequently used to improve these headlines in order to grab our attention. Many video games and mobile games are designed to become addictive using this technique and others. The newest form of human reliance is tech addiction.
However, since software is already adept at focusing attention and initiating specific activities in people, perhaps we can come up with another application for it. If handled properly, this might develop into a chance to influence society’s behavior in a more positive direction. Inappropriate use, however, could be harmful.
4. Premeditated idiocy. How can we avoid making errors?
Learning is the foundation of intelligence, whether you’re a machine or a human. In order to “learn” to recognize the proper patterns and respond to their input, systems typically go through a training phase. A system can move into the test phase when it has been fully trained, where it is presented with additional examples and its performance is evaluated.
Naturally, not every scenario that a system can encounter in the actual world can be covered during the training process. In ways that humans wouldn’t be able to, these systems can be deceived. A machine might “see” objects that aren’t there, for instance, if random dot patterns are used. If we rely on AI to usher in a new era of work, security, and efficiency, we need to make sure the machine works as intended and that individuals are unable to overcome it and use it for their own purposes.
5. Robotic racists. How can bias in AI be eliminated?
Despite having processing power and a pace much above that of humans, artificial intelligence cannot always be relied upon to be impartial and fair. When it comes to artificial intelligence, Google and its parent firm Alphabet are among the leaders, as demonstrated by the Google Photos service, which uses AI to recognize people, objects, and scenes. However, things can go awry, like when a camera failed to detect racial bias or when a program intended to foretell criminal activity displayed bias towards black individuals.
We must keep in mind that humans, who are prone to prejudice and bias, are the ones who design AI systems. Once more, if applied correctly or by people who value social advancement, artificial intelligence has the potential to be a force for good.
6. Protection. How can we defend AI from threats?
The potential for technology to be utilized for evil as well as good increases with increasing power. This is true for AI systems that, when used maliciously, can harm humans as well as robots created to replace human soldiers or autonomous weaponry. Cybersecurity will become progressively more crucial when these battles go outside the traditional battlefield. After all, we’re working with a system that is orders of magnitude faster and more powerful than humans.
7. The bad genies. How might unforeseen consequences be avoided?
We have to be cautious not merely of our enemies. What would happen if machine learning turned against us? This does not imply that the AI will become “evil” in the sense that a person might or in the way that Hollywood catastrophe movies portray AI. We may instead think of a sophisticated AI system as a “genie in a bottle” that can grant wishes but with catastrophic unintended effects.
In the case of a computer, malicious intent is unlikely to be present; instead, the wish may have been made without the complete context being understood. Imagine asking an AI system to remove cancer from the planet. After a great deal of computation, it produces a formula that does, in fact, put an end to cancer by annihilating everyone on the earth. Although the computer would have accomplished its objective of “no more cancer” quite effectively, it would not have done so in the manner that humans had intended.
8. Uniqueness. How can we maintain control over a sophisticated intelligent system?
Humans are at the top of the food chain for reasons other than having strong teeth or muscles. The intelligence and creativity of humans are largely responsible for our ascendancy. We can defeat larger, faster, and stronger animals because we can make and employ instruments to control them, including both physical tools like cages and weapons and cognitive skills like conditioning and training.
This raises a significant question about AI: Will it ever possess the same edge over us? We cannot simply “pull the plug,” as a sufficiently sophisticated machine may foresee this action and defend itself. When humans are no longer the most intellectual creatures on earth, this is what some refer to as the “Uniqueness”
9. Rights of robots. How should AI be treated humanely, exactly?
We know more about the fundamental processes of reward and aversion, even if neuroscientists are still trying to discover the mysteries of conscious experience. Even the simplest creatures share similar mechanisms with us. In artificial intelligence systems, we are developing processes of reward and aversion that are quite comparable. Rewarding increased performance, like in the case of training a dog, is an illustration of reinforcement learning.
These systems are now rather simple, but they are evolving to become more intricate and realistic. When a system’s reward mechanisms provide it with negative input, may we say that it is in pain? Additionally, so-called “genetic” algorithms function by simultaneously constructing a large number of instances of a system, of which only the most successful “survive” and merge to generate the following generation of instances. This occurs across several generations and helps a system get better. The unsuccessful attempts are eliminated. When may genetic algorithms be seen as a type of mass murder?
When we think of robots as beings with the capacity for perception, emotion, and action, it’s not a big leap to wonder about their legal standing. Do they deserve to be regarded the same as intelligent animals? Will we take the misery of “feeling” machines into account?
In certain ethical dilemmas, minimizing suffering is at stake, while in others, risking unfavorable consequences is. We should keep in mind that, overall, this technological advancement results in better lives for everyone as we weigh these hazards. We must responsibly implement artificial intelligence if we are to realize its enormous promise.