What is the most important thing to consider with artificial intelligence? Some people want to figure out the best way to make AI work for our needs. Maybe researchers want to process large amounts of data, logistics firms want to instantaneously identify the most optimal routes, or manufacturers want to maximize profits and eliminate waste. Artificial intelligence delivers in all of these areas. However, some people would argue that the most important thing to consider with AI isn’t how to make it work for us. Instead, it’s how to make sure that AI doesn’t work against us.
Does AI need a moral code?
Some experts suggest that artificially intelligent systems need a set of ethical rules to help keep them on the right track — a set of guidelines based on human morals and ethics that would shape AI decision making.
The Organisation for Economic Co-operation and Development established a set of guidelines for the development of responsible and trustworthy AI. These guidelines promote five key principles:
- inclusive growth, sustainable development, and well-being
- human-centered values and fairness
- transparency and explainability
- robustness, security, and safety
One concern is that artificial intelligence could one day be uncontrollable. The science fiction genre has shown us several possibilities for what might happen if machines go rogue. A robot revolt doesn’t seem likely. However, a network of connected devices communicating and working together — the Internet of Things — that’s beyond human control would benefit from a sense of right and wrong.
A less fictional, and far more practical, concern is the misuse of artificial intelligence. Hackers can use artificial intelligence to compromise security in a variety of ways. Equipping AI with a moral code may not prevent malicious use of these systems, but it creates an extra safeguard.
How do we choose ethical guidelines?
Establishing the rules of right and wrong might be easy for an individual. But put 10 people in a room to write up ethical guidelines and you immediately complicate things. Now try defining a morale code with representatives from different countries, cultures, and backgrounds.
Programming AI with a code of ethics is going to be a complicated and difficult process. Identifying these guidelines is just the first hurdle to overcome.