The Ethics of Generative AI: What IT Leaders Need to Know

by IT Executives Council | Jun 3, 2023 | AI

Generative AI has finally emerged as a viable technology that helps IT leaders transform their business models. The technology considerably improves the way organizations deliver content for both internal and external customers. The enhanced delivery of content boosts productivity, while at the same time improving the interactions organizations have with their customers.

As with other types of AI, generative AI is linked to several ethical issues such as copyright infringement, bias in AI algorithms, and replacing the human element in the workplace. Poor communication that leads to a lack of transparency and the risk of sharing misinformation with vital business partners also raise ethical issues surrounding the deployment of generative AI technology.

“Many of the risks posed by generative AI are enhanced and more concerning than others,” says Tad Roselund, who is the managing director and senior partner at consultancy BCG. The risks associated with generative AI require businesses to implement a comprehensive approach, which includes defining a clear strategy and making a commitment to responsibly use the technology. An organizational culture that leverages the power of generative AI for business innovation and transformation must consider six important ethical issues.

Prevent the Distribution of Harmful Content

Generative AI systems produce content based on the text prompts made by humans. “These systems can generate enormous productivity improvements, but they can also be used for harm, either intentional or unintentional,” said Bret Greenstein, partner, cloud and digital analytics insights, at professional services consultancy PwC. For instance, an AI-produced email sent from an employee might accidentally contain offensive language or explain a harmful policy change to employees.

Fake Media

The machines used for enhancing content can produce fake media, such as falsely created images, audio, and video. Unless your organization has a digital forensic specialist on staff, fake media is often difficult to distinguish from real media. The release of fake media can permanently damage an organization by reducing credibility, or worse, attracting the interest of the legal system for attempting to defame another party. Another ethical concern concerning fake media stems from using false media to harass competitors. Implementing a zero-trust framework for identity management can reduce or even eliminate the fake media threat.

Copyright Violations

Another ethical concern surrounding generative AI is the confusion associated with content ownership. The most popular generative AI tools retrieve content from massively large databases that collect information from many sources, one of which is the Internet. When a generative AI tool develops images or creates a line of computer code, the source of the data might not be known. “Companies must look to validate outputs from the models,” Roselund recommends, “until legal precedents provide clarity around IP and copyright challenges.”

Lower Employee Morale

Because generative AI possesses the capability to produce more in less time, worker displacement can become a major issue for employers. The direct connection between employee morale and worker productivity might make generative AI a counterproductive tool for organizations in search of increasing sales, while at the same time reducing operating costs. “The truly existential ethical challenge for adoption of generative AI is its impact on organizational design, work, and ultimately on individual workers,” emphasizes Nick Kramer, vice president of applied solutions at consultancy SSA & Company. “This will not only minimize the negative impacts, but it will also prepare the companies for growth.”

Lack of Data Governance

Generative AI tools access, analyze, and work with vast volumes of data that are not properly vetted for accuracy. Lack of data governance can lead to one or more employees using generative AI-produced data to unduly influence business decisions. Chief analytics officer at FICO, Scott Zoldi, explains. “The accuracy of a generative AI system depends on the corpus of data it uses and its provenance. ChatGPT-4 is mining the internet for data, and a lot of it is truly garbage, presenting a basic accuracy problem on answers to questions to which we don’t know the answer.” FICO has used generative AI tools for more than a decade to mimic edge cases for training employees to detect credit card fraud.

Bias in AI Algorithms

The tools used to produce generative AI content can magnify the negative consequences of built-in organizational biases. For example, large generative AI language models allow for human-created text and speech. Recent studies indicate the larger and more complex the human-generated content, the more likely the data collected reflects underlying social and political biases in the decision-making process. Biases in AI algorithms not only can lead to poor decisions, but also irrevocably damage an organization’s sterling business reputation.

Benefit from Generative AI With Caution

Generative AI tools benefit organizations in several ways, with the positive financial results of automatically producing content sitting at the top of the benefits list. However, when used improperly, generative AI tools can cause a wide variety of ethical issues that result in losing customers and turning off potential patrons of a business. Organizations that reap the many benefits of generative AI should dedicate a team of digital forensic experts to detect data inaccuracies, as well as prevent the fraud that develops because of data misuse.