Just because we can do something with artificial intelligence doesn’t mean we should.
42 percent of businesses are using artificial intelligence (AI) and another 40 percent are exploring uses, reports IBM. Underestimating the value of AI and delaying its adoption can pose an existential risk for many organizations. Likewise, adopting it too quickly without paying adequate attention to issues such as fairness, privacy and accountability risks unforeseeable negative consequences. Rather than an afterthought, the journey toward broad adoption of AI must include ethical AI: the implementation and management of AI solutions that prioritize responsible uses and reduced risks for all stakeholders.
Preparing for revolutionary change
The rapid advancements in AI and its transformational impact on all aspects of business operations are nothing short of revolutionary. The remarkable speed at which AI is developing has all the characteristics of a change without historical parallel. Much like recent transformational technologies such as the Internet and smartphones, AI will join and perhaps surpass them as a new context for reinvention.
There are many reasons to get excited, but poor decisions at the start of any adoption can potentially derail efforts or, worse, be a destructive force for both the organization and society. Producing results without rules or incorrect rules, for example, can create reputational risks and embarrassing results.
Specifically, generative AI relies heavily on large amounts of historical data harvested from the public web and other accessible repositories. This data coupled with algorithms shapes what AI produces. If one or both are mismanaged, the negative consequences are all too real.
Suddenly good intentions produce new and unexpected problems fairness, discrimination, trust, confidentiality, transparency, etc. AI is increasing complex questions about its appropriate use many of which remain largely unanswered. Proactively mitigating undesirable outcomes, including legal risks, requires an ethical approach to AI.
The role of ethical AI
The goal of Ethical AI is to support the responsible design and development of AI solutions that protect individuals, groups, and society from harm. Success requires a combination of efforts including IT, policy implementation and governance. It is important that this is a leadership priority.
Organizations will be best served by treating ethical AI not as a benefit, but rather as an essential part of their AI efforts. Diligence can help protect the business, enable greater confidence in deployments, and contribute to compliance with existing or future requirements. In fact, an increasing volume of national and international laws and regulations will make aspects of ethical AI required for organizations. Business leaders must realize their current obligations and be aware of developments as they emerge.
Steps Organizations Can Take Now
For organizations just beginning their AI journey, knowing where to start with ethical AI can seem difficult. This can be seen as an unwelcome effort that will slow or even discourage innovation. Despite these concerns, the benefits of ethical AI will ultimately far outweigh them.
It is never too early to start implementing an approach to ethical AI and this includes those who are already involved in their AI developments and deployments. Here are three ways all types of businesses can start increasing their responsible use of AI.
- Establish principles and standards for AI: Identify what is most important to the organization, customers, products and services, and use these to guide decisions.
- Implement AI policies and procedures: Clearly describe how the organization will implement and enforce its principles and standards.
- Create AI governance capacity: Define how the organization will provide ongoing monitoring of AI activities, including monitoring efforts, training, and continuous improvement.
For too many people, implementing ethical AI may not be a top priority today, but it needs to be. Leaders need to be convinced that doing the right thing with AI will get the right results.
Creating and implementing a viable approach to ethical AI offers the best chance of avoiding significant problems. Without ethical AI, organizations may think they are on the fast track only to wake up to find they are operating on the wrong path.
Special thanks to Noah benShea for his contribution to this article.
