One of the most common objections to AI is that it is “unethical” to use it in business. This can be taken two ways. First, if the company is not doing anything illegal, is it being unethical? And second, if they are unethical, what is it that they are doing?
The first answer to the question is both yes and no. If a company is not breaking any laws, using ethically correct data, and doing things in the public’s best interest, it is not unethical. However, this is a rather vague answer, as many gray areas surround business and human interaction in general. Let’s take a closer look at some situations where a company might be asked to use ethical guidelines.
One popular area is in product and service design:
The company may have a vision, but that vision may be difficult to realize. It may require exploiting resources to realize that vision. In this situation, the company is being asked to do what is ethical but creates fewer short-term problems and long-term value.
Another area is machine learning:
Here, the company may wish to teach an autonomous system how to recognize a particular image or word instead of teaching a human being how to recognize a particular item. Again, it could be viewed as whether or not the solution created more significant benefits than the problem being solved. Of course, the company could be viewed as intentionally using a less than desirable result to help the world, but is that worth it?
AI ethics is also closely tied to issues of access and privacy:
As computer technology becomes more developed, companies may be held accountable for the information they collect about their customers’ private lives. Privacy is also an important consideration. How does a company decide what is private and what is not? This is particularly tricky for artificial intelligence software because it may be unclear which behaviors are appropriate.
There is one area that the company may choose to skirt around, however. If the company’s mission is to develop products that can replace human doctors and nurses, then it is legally allowed to develop a computer program that can do those tasks. However, it must be clearly stated that the product is to replace a doctor or a nurse. Otherwise, it is illegal to use this product to provide health care to anyone. Companies may claim a slight loss for this purpose, but they must clarify that the doctor-nurse relationship was always intended to exist.
One of the most controversial areas of the discussion has been artificially intelligent software to decide what is ethical and what is not. Companies have been accused of deciding what good moral behavior is and what is not. Proponents argue that an artificially intelligent computer cannot be manipulated in this way. However, many people are concerned that if a computer cannot be manipulated, it will be completely free to choose whatever path it chooses.
The company has had to make some changes to its current practices in this area. It now stipulates that all employees will need to sign two separate agreements before sharing any information from the company. Each employee will agree that the information is not to be used by other companies for commercial purposes.
Additionally, it states that all employees will need to sign an additional agreement that states that the company will not use their information for commercial purposes without first receiving written permission. All of these changes have increased employee trust in the company and increased employee morale.