Humanizing Generative AI
January 23, 2025
“Evidence-based leader selection and development offers significant benefits to organizations, enhancing their overall performance and success. By leveraging data-driven approaches, companies can make more informed decisions about who to place in leadership roles and how to nurture their growth. One of the primary advantages of evidence-based practices in leadership is the improvement of decision-making processes. This approach challenges assumptions and personal biases, leading to more objective and unbiased decisions, as well as more complex data analysis that can guide more nuanced decisions.
By relying on empirical data and proven methodologies, organizations can select leaders who are more likely to succeed in their roles, ultimately contributing to better organizational outcomes. Evidence-based leadership practices also contribute to improved organizational performance. Companies that invest in this approach can expect increased efficiency, effectiveness, and overall performance.
From our direct experience working with a large restaurant chain, we leveraged behavioral tagging (one type of evidence-based tool) to gain insight into the performance of new leaders. This approach allowed us to identify specific behaviors that correlated with successful leadership in the restaurant environment. By observing and tagging behaviors with a bespoke Ideal Employee Model, we were able to create a more accurate profile of high-performing leaders. This evidence-based approach enabled us to refine our leader selection process, focusing on candidates who exhibited behaviors associated with success in our specific context.”
What did you think? The above article was written by Perplexity, a generative AI tool. My prompt was quite generic, providing just the basic information about the topic and an example from our work. The tool did not provide a bad description of the topic…but below are a few questions that I would ask:
1. Did you find the writing style engaging? Did it leave you wanting to learn more?
2. If you had no background in the work, did the content offer sufficient insight?
3. Were you convinced of the value of evidence-based leader development?
I certainly could have improved the prompt to make the article stronger, but, that said, I’m seeing a lot of AI-generated material of this nature. Many efforts to improve AI efficiency have huge trade-offs in value and impact of the writing.
I believe generative AI is offering a great deal of value, allowing strong users to focus on the effectiveness and impact of their writing; generative AI also supports those who have limited writing skills but nonetheless have something interesting to say. But to bring out the maximum value in communication, the human touch must be re-introduced into the writing (whether in better prompts or in significant editing).
If you are starting down the generative AI path, here are a couple suggestions and cautions:
1. Test out different tools. The prompt used for the above article produced different results in ChatGPT and Gemini. Find your preferred tool (I like Perplexity because it offers source links to its resources.)
2. Learn proper prompting. Develop skills in crafting effective prompts. The way you phrase your instructions significantly impacts the quality of the AI output.
3. Prioritize data privacy. Remember, these are public data sets you’re sourcing and to which you’re contributing! Be cautious about sharing sensitive or personal information with AI tools. Avoid inputting confidential data or personally identifiable information.
4. Verify and fact-check. Always double-check the accuracy of AI-generated content. Generative AI can sometimes produce incorrect or biased information.
5. Use AI as a complement. Leverage AI to augment your skills and knowledge — not to replace them. Combine AI outputs with your own expertise and judgment, and give the product your unique touch.
6. Practice responsible use. Be mindful of potential risks and ethical considerations. Use AI tools in a way that respects intellectual property rights and avoids harmful applications.
I hope this provides a useful start. And yes, I did use a generative AI tool to augment this list. 😊
Humanizing Generative AI was originally published in Horizon Performance on Medium, where people are continuing the conversation by highlighting and responding to this story.
