Machine learning is not a new tech development, but the ethical issues artificial intelligence presents are at their most historically pressing moment. There are multiple ethical questions to confront, including the four most prevalent below.
1. Privacy Is The End Of An Era
Although some government regulations ensure consumer privacy (like the GDPR), the potential for A.I. to understand buyers on an intimate level is enormous. Some potential examples of privacy breach include:
- Data Exploitation: The collection of our data from consumer devices (like cell phones or blue-tooth enabled light bulbs) to provide personalized service.
- Identification and Tracking: Tracking is great for a lost iPhone, but less great for individuals who need to disappear, like domestic abuse survivors.
- Voice and Facial Recognition: Facial recognition could be used to track and control minorities, or encourage the occurrence of corrupt police activity, even though we think of facial recognition now as an easy way to unluck our iPhones or try on glasses virtually before ordering them. Additionally, we are already seeing issues with voice recognition technology, as tech giants like Amazon have admitted to listening in on.
2. Robots Will Rule Humans
What is a human if not defined by autonomy, and should we give that right to artificial intelligence?
According to Nick Bostrom, a superintelligence expert at University of Oxford, it’s possible that the malicious robot overlords of our imaginations could be a reality if future A.I. systems hold values that don’t align with human values. According to Ray Kurzweil, futurist and Chief Engineer at Google, this future (i.e. singularity) is closer than we think. He perceives this future as something to look forward to, though, not fear.
3. A.I. Will Take Over The Job Market
Job security is always the initial scare tied to advanced technology. This is a logical fear, since work is directly tied to compensation, which is tied to survival. There are two possible outcomes of economic inequality and labor rights issues at the hands of artificial intelligence.
- Inhumane Practices: Human workers (at least in the context of blue-collar laborers) could be replaced by advanced robotics that reduce the risk of dangerous jobs. However, there is a fear that A.I. would implement further labor on its human counterparts, or drive human workers into an early grave.
- Contributions to Inequality: If A.I. takes over human labor, how will unskilled workers earn a living wage? Although one answer is to implement universal income, that policy is sometimes perceived as a socialist tactic. Instead, A.I. could be a catalyst for a spiral of economic inequality.
4. Moral A.I. Might Not Be Possible
Despite our best human efforts, it may be possible that A.I. will never be moral. The issues with training A.I. systems to act ethically include:
- assuming A.I. can be motivated to act ethically
- expecting a universal ethics to emerge
- believing the humans training A.I. to be ethical are themselves ethical
Experts like Stephen Hawking, Bill Gates, and Elon Musk have all stated that superintelligence should be handled delicately.