What’s next in KM: All roads lead to AI
The U.S. has published a Bill of Rights for AI, focusing on safe and effective systems, discrimination protection in algorithms, privacy, informing users about the system they are using, and being able to opt out of automated systems. Currently, these are guidelines, but legislation in the future is likely. The absence of AI-specific legislation does not negate the existing requirements for privacy, transparency, and accountability that are in place, however.
For now, governance of AI is driven not only by the desire to protect a company’s brand but also by enforcement of consumer protection regulations. If customers believe that the results of an AI process are suspect due to bias or lack of transparency, the organization’s credibility is undermined. In addition, it puts development costs at risk. “Companies are concerned about fines but also about the losses from having to delete the algorithm that runs their model,” Brannon explained.
In 2023, the Federal Trade Commission (FTC) reached a settlement with Rite Aid, which had used facial recognition to identify suspected shoplifters. The company was required to delete any images of consumers collected via this surveillance, along with any algorithms developed using such images. Consumers must be informed if biometric information is being processed or action is being taken against them resulting from the company’s use of the system. Organizations can expect increasing scrutiny in their governance practices as AI is integrated with their KM systems.
Increasingly social
Social media has a role in several aspects of KM, but the dominant one is in marketing. Social media provides a unique way to connect with individuals and is a major way that people learn about products in today’s markets. It is also a channel through which customers provide comments to others within the social media ecosystem. Data from analyzing social media interactions is used to refine outreach messaging to customers and to provide feedback for strategic planning.
Many organizations do not have staff with the expertise in social media marketing, however. At least half are estimated to outsource this function. Strella Social Media has been providing social media marketing services for companies across multiple verticals for more than 10 years, and within the last year, it has begun using ChatGPT for audience research and to generate initial social media marketing messages.
“The most important step for a company to take is to understand its audience,” said Laney Goff, executive VP at Strella. “We can then help with organic growth and connecting a brand with its audience. We use ChatGPT to help us understand what the audience wants. It looks through any content pertaining to the topic we are researching, including blogs, white papers, and websites.”
Refining the ChatGPT prompt to the proper focus is critical. “We might tell it to act as an expert social media person or a marketing manager, and then instruct it to tell us the frustrations of people who use a certain product, and identify other features of that target audience,” she continued. “The more specific our instructions, the more accurate the response.”
Strella is also using ChatGPT to generate ideas for posts, but it has some caveats. For example, ChatGPT-generated content for posts should be considered as a draft, not a final. It should be filtered for bias and personalized for the brand so it does not sound generic. No AI tool has the subtlety or nuances of a human, nor a sense of humor, so there should always be an overlay of refinement by having a human-in-the-loop.