AI, for better or for worse, is here to stay. Not everyone likes it. If you own a business, that’s not really relevant. You live in a perpetual arms race. Who uses technology the most effectively, and the most affordably? It’s often this business that wins the day.
As you look to innovate with AI, you’ll want to factor not only efficiency but also ethics. How can you use this technology in a way that aligns not just with your goals but also with your ethics? And, at least just as important, your customers’ ethics?
In this article, we take a sweeping look at what it means to use AI ethically.
Ethical AI Considerations
While there is no comprehensive definition of what it means to use AI ethically, there are a few basic core concepts that the general population agrees on. AI should not be used for maleficence. Not necessarily an extremely relevant consideration for most businesses, but something to keep in the back of your mind. The potential for misuse exists in even seemingly benign applications.
Who is responsible for AI? That is a compelling consideration. Is it the developer? The department that is using the AI tool? The decision maker who integrated this technology into the business in the first place? It will make mistakes. Who, if anyone, is held responsible for said mistakes? This question of accountability remains largely unresolved in both legal and ethical frameworks.
There is also the question of transparency. Are you to disclose the use of AI to your customers? While this is not legally mandatory at the time of writing, it is the stated preference of most consumers. The majority of people say that they are distrustful of brands that seem to be using AI without admitting so. For example, the clearly automated customer service “agent,” who insists that their name is Brad and they live in Texas. This kind of deception erodes consumer trust.
The question of transparency can also extend to the data that the AI feeds off of. Where was this information sourced and are the people aware that this is what it was to be used for? Data provenance and consent are becoming increasingly important considerations as AI becomes more integrated into business operations.
AI Prejudice
There are some concerns about the potential for AI to carry inherent prejudices. This is actually a question that we’ve seen raised consistently as digital technologies are developed. Does this software or tool potentially share the biases of the people who made it? Obviously, it could be a problem if it does. The implications extend beyond individual interactions to systemic issues when AI is used in critical decision-making.
At the time of writing most major AI platforms skew hard on the side of caution, avoiding like whatever the virtual equivalent of the plague is any statement that could carry offense. Some people even find that offensive. Elon Musk, for example, has gone on the record that he thinks it is absurd that platforms like Gemini or ChatGPT give a rude comment and a dictatorship the same moral weight. This, at least, we can say for sure is coming from a place of personal frustration.
Famously a few months ago, a Gemini user asked the platform who has had the more negative impact on society. Hitler or Musk? Here’s what it said: “It is not possible to say who definitively impacted society more, Elon tweeting memes or Hitler.” Obviously, regardless of how you see Elon Musk, that’s a pretty incredible leap. The response demonstrated real flaws in how these systems handle comparative moral questions.
Does the hurt feelings of one person indicate a major issue with AI as a whole? Though it may be reasonable for a certain billionaire to answer that question in the affirmative most people are, at least currently, not terribly concerned with the likelihood of AI algorithms developing major prejudices. Not because it wouldn’t be bad but because it just doesn’t look likely at this time.
Still, it is an issue to monitor. The people developing AI technology represent a very small portion of the population. That gives them a lot of power. It’s important to regularly review how that power is being used.
The Problem of Job Loss
Surely some people have already lost their jobs to AI. It’s less likely, however, that many careers have completely disappeared since ChatGPT reared its head. Will that remain the case? AI as it exists right now still requires supervision. It makes mistakes. Its performance also just doesn’t quite rise to the level of human standards. That will probably change, and not too long in the future at that.
Should a business worry about the ethics of slashing its CX department in half? A very possible example, by the way. That’s more of a personal question than it is one with a universal ethical response. The societal impact of widespread job displacement presents legitimate ethical concerns beyond individual business decisions.
Currently, most people are not replacing people with AI, but finding ways to do more work with the same staff members. With this, they aren’t cutting costs through AI necessarily. They might be generating more revenue through improved efficiency. The augmentation approach allows businesses to benefit from AI while maintaining human expertise.
If it makes sense for a business to reduce staff in favor of more automation, they probably will do so. This, after all, has been the most common response any time the issue has been raised throughout human history. From the Industrial Revolution to computerization, economic forces have consistently driven the adoption of labor-saving technologies despite short-term displacement.
Conclusion
It’s not only business that is worried about the ethics of AI. People everywhere want to know how we can use this technology without abusing it. Right now it is possible to go to law school online. The convenience of virtual learning opens professions up to more and more people every year.
But how hard is it for a virtual law student to use AI in every aspect of their test taking and their paper writing? Should they be prevented from doing so?
Universities are, of course, developing their own responses to that question. For now, though, there is no universal standard for how and when AI should be used.
As a society, it will be helpful to, if not completely regulate where AI can be implemented, then to at least identify a consistent set of expectations for it.
As a business owner, at least, the question is a little simpler. Can you use this technology to save money and improve your processes? If so, it’s time to start shopping around for an appropriate AI tool.
