By Elyssa El-Hajj
Generative artificial intelligence (Gen AI) is becoming increasingly more common and normalized. Surveys indicate up to 40% of the nation’s workforce currently utilizes AI. This number will only increase as more organizations shift to embracing Gen AI. This is how to become more efficient and save time and energy.
Everyone experiences the repetitive or routine tasks that happen day to day. They take time away from larger or more complex projects. Things like filling in paperwork, summarizing meeting notes, and looking up information can easily be streamlined using Gen AI. There are infinite ways of implementing AI in the workplace. My concern with the never-ending possibilities of AI is that we won’t slow down to consider issues like misinformation, privacy violations and inauthenticity.
To understand the pitfalls of Gen AI, first we need to understand how the model is built and where all the information comes from. Gen AI functions based on the prompts from end users. Then, the technology will source answers from all the data it has access to, which is a lot both in sheer amount and in range of reliability. AI models train on sources like public data sets that pull from books, scientific journals, news sites, and free image libraries – all the avenues one might look at for information. However, Gen AI also trains on sources like user-generated content (UGC), pulling from social media, forums, and other sites where there is little oversight into what is uploaded. Here is where we run into issues. How often does one come across misinformation or opposing opinions while scrolling through the web? I could go and change an entry on Wikipedia right now, that won’t make it accurate or even remotely true, but Gen AI will still take it into their training model. This is a crack in the system where inaccurate information can be the output.
AI trains on the information it’s given, and that can be a problem. Also, the risk of violating privacy is concerning. Consider the example of an individual generating a report internal to the organization. To save some time, they decide to input the report into a Gen AI to create a summary. That internal report could have contained information that wasn’t made public. It may even contain private identification data on a worker or customer. All that data is now available to AI and can be used in future data sets.
Misinformation and privacy violations are serious outcomes while using Gen AI but don’t happen every day. What does happen when using Gen AI daily? Most concerningly is the loss of authenticity and personal voice. While Gen AI can easily be used to fix grammatical errors, it is just as easy to generate a response to an email or a direct message. When the reply is a simple “Thank you!” or “I will have [task] done by the end of the day.” there is no issue. However, when responding to a customer concern, an AI generated response might fall flat on providing a personal touch. This will affect the quality of services and solutions. Much of customer service is being able to communicate with the individual in a way they understand and feel heard.
AI currently does not have the same context on the individual as the service team does, especially if there is a previously established connection. A service team member is better equipped to explain a technical issue in a way that the customer can understand compared to an AI. Its output can be overly complicated and not specified to the issue at hand. Building a connected and collaborative organization cannot be done with 100% AI generated emails being sent back and forth. This lack of sincerity also extends to connecting with the public. News sites and article publishers have noted that pieces generated by AI receive significantly less viewership. Gen AI can scrub your voice from the dialogue, flattening the tone and taking away unique expressions that are specific to the individual. Gen AI has the ability to sound human at times, but it cannot replace the human touch in writing. That is what generates emotion, tone, and storytelling which makes even scientific journals more interesting to read.
Don’t get me wrong, I use AI, in fact I used it to assist in the creation of this article. However, I’ve avoided the pitfalls. See, Gen AI is developing new abilities every day but understand that Gen AI does not replace all the functions of an individual. Use this new technology. Use it while placing an emphasis on education of its use. Discuss its limitations and consider what data sets are being used. See if there are any privacy safeguards put in place. Treat Gen AI as a first draft and fact check your information. Always edit the output to fit the needs and preserve your voice to increase authenticity. Everyone is adapting to the new age of AI, pushing us to higher levels of productivity at lower effort. It is exciting to see how the tool develops; however, it is still just a tool. As such, please consider the previous advice on making AI a guide, not a crutch in doing office work. Your co-workers and customers will all see the difference if it is used correctly.
For something fun to think about, can you tell where AI was utilized in this article? If you can’t, that was the point. Challenge yourself to work in synchronization with AI.
