What are ethical and compliance concerns of AI?
September 3, 2019
It’s impossible to ignore the growing influence of artificial intelligence (AI) on our personal and work lives. In fact, 20 percent of C-level executives attest to using machine learning as a core part of their business.
However, with increasing capabilities comes more exposure to corporate compliance risks. As AI becomes more woven into daily work, it opens new ethical and legal concerns. In order to avoid compliance violations, there are ethical considerations corporations must actively work into their increasingly automated worlds.
Dangers of Discrimination and Bias
It’s easy to forget that AI is built by humans. That means the technology is susceptible to the biases and prejudices of the humans who create it. Because humans are building software and algorithms that feed AI, machines are vulnerable to prejudiced programming and can spit out information based on unconscious biases.
In many cases, discrimination is unintentional. However, it is important to remember that AI output for decision-making is based on the data that goes into creating it. For instance, one high-profile company recently built AI with the intention of screening job candidates fairly. However, in action, the AI showed favoritism toward male resumes. The cause? The machine’s algorithm was set up to vet candidates based on the tech company’s hiring patterns over the previous decade—a time period when men overwhelmed the tech market.
In some cases, even when potentially discriminatory categories such as gender and race are excluded from algorithms, AI output can be compromised if the underlying data is the result of historically biased or discriminatory systems. For instance, imagine a company designs AI that ignores race or gender, but focuses on contacts’ addresses. In the U.S., where neighborhoods are often largely segregated by race, culture, and socioeconomic conditions, the company could be using AI to unintentionally target specific groups.
To avoid practicing discrimination and violating laws, businesses should monitor AI at every stage—from development and testing to operation and conducting audits. At every turn, they need to make sure their machines are promoting fair, ethical practices and are free from potential violations.
Data Security and Data Privacy Concerns
Data is what feeds AI, and that can often involve the storage and use of large amounts of data. Along with data collection and use come data security and privacy concerns.
In the area of data security, organizations must take steps to safeguard the data that’s being used for AI. Organizations should make sure they’re implementing meaningful training to build a culture of security awareness. This will help protect against breaches that could compromise the valuable data of customers, employees, and the company.
With regards to data privacy, organizations must ensure they are familiar and complying with data privacy requirements when it comes to collecting and using individuals’ data for AI. This includes giving individuals proper notice and obtaining appropriate consent before gathering or using data.
Control and Security Concerns of AI
AI means a new breadth and depth of efficiency. Unfortunately, it could also mean new opportunities for misappropriation and criminal action. Companies need to safeguard against these dangers.
In the future, responsible companies should make sure AI is being developed in a way that’s consistent with their corporate values and internal policies. They also should ensure AI doesn’t fall into the hands of bad actors. That means building a culture of safety and undergoing meaningful security training as your company develops and uses AI.
Antitrust Capabilities with AI
When used unethically, AI can allow corporations to make decisions that corner the market and eliminate competition. This makes it a potential antitrust liability. For instance, consider a scenario in which a large company decides to set up AI to collaborate on pricing decisions—undercutting competition, and driving other companies out of business.
As regulators become increasingly alert to these marketplace dangers, companies should pay special attention to their actions. They should take care to avoid developing AI with intentions of justifying anti-competitive decisions. Regardless of intentions, it is important that your company isn’t developing algorithms that violate antitrust laws.
Emerging Corporate Compliance
As AI is becoming more integrated into company procedures, laws and regulations are emerging in turn. Government bodies are zoning in on the antitrust possibilities AI could encourage. On the federal level, from executive orders to proposed legislation, lawmakers are diving into the potential uses and abuses of AI. Eventually, this emphasis will result in new compliance requirements throughout the public sector, private sector, and international spaces.
In order to navigate this evolving compliance landscape, companies should ask themselves how they’re programming AI and how they’re protecting against potential risks that could come in its wake. Luckily, there are resources to help. Companies can implement training from established partners that helps employees recognize bias and other potential issues. Professional partners can help you avoid ethical and compliance pitfalls as you participate in the ongoing AI revolution.
Got a learning problem to solve?
Get in touch to discover how we can help