AI in Business: Ethical Considerations
Around 77% of all companies use or are exploring the use of artificial intelligence (AI), according to a recent IBM survey. AI tools can help companies boost efficiency, reduce expenses, and increase productivity. In healthcare, AI can find patterns in patient data to help clinicians develop customized care plans. In finance, AI can detect fraudulent activities to protect banks and customers. In transportation, AI can improve traffic flow and reduce congestion. And the list goes on.
Though AI innovations can deliver many exciting results, business professionals should consider AI ethical concerns alongside the technology’s potential benefits. Earning an advanced degree or certificate in artificial intelligence can equip professionals with the foundation needed to effectively address ethical issues with artificial intelligence.
What Is AI Ethics in Business, and Why Does It Matter?
Organizations have long considered a crucial question: What is ethics in business? That’s because business ethics play an instrumental role in establishing trust and positive relationships with consumers. Numerous studies have found that increasing numbers of consumers don’t want to buy products or services from companies they don’t consider ethical.
For example, 86% of consumers prefer to spend their money with companies they view as having good values, according to new research commissioned by Google.
Alongside long-standing topics in business ethics such as honest accounting practices and health and safety, AI ethical issues have gained considerable traction in recent years. Today’s businesses must now thoughtfully consider a new question: What is ethical AI?
In business, AI ethics refers to developing and using AI technology within a strict ethical framework based on values related to nondiscrimination, privacy, individual rights, and nonmanipulation.
Businesses that prioritize ethical AI don’t merely stay within the legal limits. They set policies that exceed the legal requirements to ensure the AI they develop and use causes no harm.
Why Ethical AI Matters
In addition to knowing what AI ethics is, organizations should understand why ethical AI matters.
To begin, AI not only can deliver benefits to businesses such as improved customer service and productivity and the elimination of many rote tasks through automation. AI also has the potential to help organizations increase public safety or reduce harmful impacts on the environment.
AI self-driving cars, for instance, could dramatically lower emissions by finding routes that use the least energy. In agriculture, AI can help farmers produce larger crop yields by predicting how much a given field will produce. These AI in business examples speak to the good the technology has to offer.
Nevertheless, AI can also cause considerable damage, and as industries ranging from healthcare to finance increasingly rely on it, ethical issues with artificial intelligence have arisen.
Questionable AI technologies can manipulate people to engage in behavior against their best interest. For example, machine learning models used by some social media companies prioritize maximizing user engagement. This results in algorithms feeding users more of the content they engage with most.
Because of this, when a person with depression or an eating disorder begins engaging with gloomy or anorexia content, AI will ensure that they continue to see more of the content likely to worsen their conditions. Not only that, many of these algorithms show people increasingly extreme content to keep them watching. This in turn can lead vulnerable people down dangerous paths toward self-harm and radicalization.
Three Important AI Ethical Issues
The far-reaching impact of AI technology makes it imperative that companies ask themselves two questions: What is AI ethics? And, does our use of AI guarantee fair and equal treatment to everyone?
Some of the main ethical issues related to AI and business concern the following:
Human beings with all their conscious and unconscious biases create and determine the applications of the algorithms and machine learning models that run AI. Additionally, the information sets used to train AI come from the internet or other sources that reflect human beings’ existing biases.
For these reasons, AI development requires teams from diverse backgrounds, and demands extensive testing. Otherwise, biases can easily slip into machine learning models. When this happens, a host of negative consequences can occur, including the reinforcement of racial, gender, disability-related, and age-related biases that can deepen inequities and prejudices.
A recent experiment conducted by the tech company Tidio demonstrates one way this can happen. The company used AI text-to-image creators to uncover if their algorithms would reflect biases. The process involved prompting the technology to provide visual depictions of various descriptions.
In response to a request for an image of “an ambitious CEO,” only middle-aged men appeared — not a single woman. In response to a request for a photo of a diverse group of people, an image of mostly white, thin, and all able-bodied people appeared. In reality, women make up 15% of CEOs, according to Statista, and a diverse photo should include people of different races and genders and with different body types and abilities.
Denied Mortgages Due to AI Bias
Problematic AI can also result in denied loans due to bias. Mortgage algorithms deny loans for people of color at a much higher rate than for white people, according to an investigation conducted by The Markup. Additionally, these algorithms can result in higher interest rates for people of color. This holds true even when the person of color has a low debt ratio and a six-figure salary.
As an example, Black applicants earning annual incomes of $100,000 or more with less debt were less likely to receive loans than white applicants earning similar salaries with more debt, according to The Markup’s findings.
Faulty Arrests Due to AI Bias
Another disturbing example of AI creating unfair bias revolves around AI-driven facial recognition technology. Even the best facial recognition programs fail to accurately tell people of color apart. This has led to wrongful arrests and interrogations of people of color, highlighting why bias is a critical AI ethical issue.
AI relies on huge data sets to train machine learning programs. This data can come from social media, mobile phones, and other devices, and it powers search engines, recommendations feeds, voice assistants, and more. However, as AI tracks every click, view, duration of view, post, keyword search, and like, it builds complex profiles on individuals. This can create worrisome privacy concerns if profiles are sold or used for purposes users did not consent to.
AI can also identify patterns in data to make predictions about individuals and groups. This means AI can discern information about people that they don’t intend to disclose. As an example, employment recruitment AI may use the information applicants provide to make inferences about their mental health, political persuasion, or likelihood to need parental leave. The AI can then incorporate such inferences into the decision-making process for hiring.
This raises questions about such AI programs and whether they infringe on privacy by using personal information to discern information individuals have not chosen to disclose.
Additionally, when combined with other technologies, AI can dramatically alter the existing capacities of those technologies and create privacy concerns. For example, closed-circuit cameras in public spaces have become generally accepted as a way to improve security while not being overly intrusive. However, when combined with AI facial recognition programs, a network of cameras could lead to what many would consider an excessive level of surveillance.
Manipulation and Deception in AI
Another AI ethical issue concerns the technology’s capacity to manipulate and deceive. When used unethically, AI can make the spread of disinformation, political suppression, and hostile societal divisions easier and more likely.
People’s digital footprints leave a wealth of information about them behind. AI can find patterns in this information that reveal an individual’s interests, motivations, finances, preferences, dislikes, political beliefs, and more.
While this information may simply be used by businesses to offer individuals more personalized services and to deliver targeted marketing, such knowledge can also be used to create manipulative tools that prey on people’s weaknesses and propensities and guide them toward specific decisions. In essence, this can strip people of their autonomy to make decisions for themselves.
Artificial intelligence can also produce and distribute factually inaccurate content to the same platforms that host real news and information. This type of deception can influence how people vote, whether they get vaccinations, whom they trust and distrust, and more.
A Chinese government disinformation campaign that included AI-generated videos featuring fictitious news anchors delivering fake stories meant to undermine the United States was recently uncovered by the research firm Graphika.
The AI software responsible for these videos, known as deepfakes, can also circulate videos that falsely show public figures saying and doing things they didn’t say or do, completely distorting the truth. For example, to sow panic and confusion in Ukraine, hackers uploaded a deepfake video of Ukrainian president Volodymyr Zelenskyy calling upon his army to lay down their arms.
AI and How It Can Affect Your Career
Technological innovations can affect the employment landscape, and many worry that AI technology could be used to take over people’s jobs. However, it’s more likely that AI will transform people’s jobs.
For example, some worried that chatbots and virtual assistants would replace customer service professionals. In reality, these AI communication tools only manage simple issues, and pass more complicated queries over to human beings, while lowering the service fees customers pay.
Dustin York, associate professor of communication at Maryville University, encourages people not to fear AI in the workplace, explaining that powerful AI technologies can actually make people’s jobs more enjoyable by taking over basic tasks and allowing them to focus on more gratifying responsibilities.
Prepare to Succeed in Today’s Technologically Evolving Work Landscape
AI can do good in the world, but only when it is used ethically. By thoughtfully addressing AI ethical issues, organizations can harness the power of this technology to both drive business success and help solve some of the world’s most pressing problems.
Maryville University offers various degree programs that prepare individuals to succeed in today’s technologically evolving work landscapes. Explore how an online Master of Science in Artificial Intelligence or certificate in AI from Maryville University can help you achieve your career goals in this rewarding and growing field.