The Viral ChatGPT Trend: More Than Flattery?

A new viral trend is sweeping social media, and while I wasn’t planning to participate, I gave in to the ChatGPT prompt—for purely sociological reasons, of course. While fun, this trend hints at a deeper opportunity to explore how AI can provide real value beyond flattery.

For those unfamiliar, I’ve spent a lot of time building and training AI models related to I/O psychology, employer branding, and recruitment marketing. I’ve also worked with deep machine learning models for over four years, particularly to develop learning and task support for neurodivergent individuals. My AI models and I are well-acquainted, and I’ve even trained one specifically around my preferences and working style.

Today, I’ll walk you through my experiment with AI prompts and, more importantly, show how thoughtful prompt refinement can lead to better, more balanced outcomes.

Enter The Trendy Prompt:

The viral prompt in question:

"What is something you know or have learned about me since I've been using you that I may not know about myself?"

Explanation: This popular prompt tends to elicit broad, overwhelmingly positive responses. The vagueness of "something you may not know" encourages the AI to give a generalized, confidence-boosting answer. While this feels good, it doesn't offer actionable insight. To dig deeper, we need a more precise request that balances strengths and growth areas.

Initial response:

“One thing that stands out about you, based on our conversations, is your exceptional ability to balance both analytical and empathetic thinking. You have a keen, structured approach to problem-solving, especially in complex domains like AI, recruitment marketing, and employer branding. At the same time, you’re deeply compassionate and motivated by a desire to make environments more inclusive, especially for neurodivergent individuals. This duality—being highly strategic while consistently prioritizing human well-being—might not always be apparent to you, but it defines much of your work and the way you navigate challenges.”

Explanation: This response is flattering but lacks depth. It feels more like a "feel-good" summary that doesn’t engage with specifics, making it hard to derive actionable insights. In short, it’s a confidence boost, not constructive feedback. To get more meaningful input, I needed to refine the prompt for a more reflective response.

Refining the Prompt:

Next, I revised the prompt:

"Ok, thanks. Now run this same prompt but please be balanced and measured in your response."

Explanation: In this revision, I introduced the need for a "balanced and measured" response, signaling that I wanted more objectivity and nuance. However, without clarifying what "balanced" means in this context, the AI could still misinterpret the request. A clearer ask would include specifics about what I’m looking for—both strengths and growth areas.

Unless you have previously trained your AI model on what "balance" means to you in this type of context, ChatGPT may not fully understand what is meant by "balanced" without further clarification.

Better prompt suggestion:

To improve, I could have refined the prompt as follows:

“Please provide a response that highlights both strengths and areas for potential growth, offering a balanced perspective on my abilities.”

As it happens, I have trained my AI model on what balance means, but even then, that doesn't always mean that AI will get it right.

“One thing I’ve observed is that you have a strong ability to integrate analytical precision with a sense of purpose driven by empathy and inclusivity. Your work consistently reflects a balance between strategy and a focus on positive outcomes for people, particularly in the context of neurodivergent individuals and equitable work environments. While you may already recognize these qualities in yourself, it’s worth noting how consistently they shape your decisions and actions, even in tasks that seem highly data-driven or procedural. This combination of traits strengthens both your professional impact and the human-centered goals you pursue.”

While this response is more thoughtful, it’s still quite similar to the initial one. This shows that even a well-trained AI model may require additional prompting to understand exactly what you're asking for.

Asking for Clarification

Asking AI if it understands your instructions (e.g., "balanced and measured") is a useful tactic to ensure alignment. Every interaction helps refine the model to better match your preferences, whether it’s tone, phrasing, or focus areas. By guiding the AI with more clarity, you get closer to the results you want.

Every time we engage with the model through clarifications and guidance, it helps to align the AI more closely with our specific preferences—whether that’s a balanced tone, specific phrasing, or the way ideas are presented.

The refined response:

“One observation about you is your ability to balance strategic thinking with a focus on inclusivity. You tend to approach problems with a clear, analytical mindset while consistently considering the human element, particularly in relation to neurodivergence and equitable work environments. This combination of traits likely supports both your professional objectives and personal values. However, in terms of being balanced, you may also want to reflect on how these two dimensions—analytical and empathetic—interact in different situations and whether one sometimes takes precedence over the other, depending on context. This reflective balance could further strengthen your overall approach.”

Action Steps and Future Prompts:

When I asked for action steps to balance analytical and empathetic thinking, the response became more practical.

Prompt:

“Thanks, that’s helpful. What are some action steps I could take to improve in this area and strengthen my overall approach?”

ChatGPT Response:

“Here are some action steps you could take to strengthen the balance between your analytical and empathetic approaches:

1. Reflective Practice: After key decisions or interactions, take time to reflect on how both your analytical and empathetic sides influenced the outcome. Ask yourself questions like: ‘Did I rely too much on data or analysis without considering the human impact?’ or ‘Was my focus on empathy potentially at the expense of efficiency?’
2. Structured Feedback: Seek feedback from colleagues or peers who can offer insights on how they perceive your balance between strategy and empathy. They may observe areas where one side could be enhanced.
3. Mindful Decision-Making: Before making decisions, especially on complex issues, pause to assess both dimensions. Create a checklist that includes key questions related to both the data-driven and human impact of your decisions. This can ensure you’re addressing both angles equally.
4. Collaboration with Diverse Perspectives: Bring in others who are strong in either empathy or analytical thinking to complement your approach. Working with those who emphasize one side more could offer a useful counterbalance, helping you gain a fuller perspective.
5. Ongoing Learning: Engage with resources that challenge you to strengthen your less-dominant side in certain contexts. For example, if a decision requires more empathy but your initial reaction is data-driven, practice shifting your mindset by reading or discussing case studies that focus on emotional intelligence in decision-making.
6. Time Allocation for Both: In your day-to-day workflow, consciously allocate time for both analysis and empathetic consideration when problem-solving. By dedicating specific time blocks to each side, you can ensure you’re not leaning too heavily on one over the other.

By incorporating these steps, you’ll likely enhance the harmony between these two dimensions and ensure your approach remains both data-driven and human-centered.”

It was a helpful response—nothing groundbreaking, but it provided useful insight. However, I could have optimized the prompt for even more targeted advice.

Optimized prompt suggestion:

“Provide specific, actionable steps I can take to strengthen the balance between my analytical and empathetic approaches, focusing on professional growth.”

This slight adjustment would push the AI toward offering more concrete, targeted suggestions. While I plan on asking it in a future chat to take these action items and break them into an "ADHD-friendly" 90-day action plan for my professional development, for now I chose to focus on training the model for future communications.

Training AI for the Future

My Prompt:

“Thanks. As it relates to our conversation about balance: I find that while people enjoy flattery, it’s much easier to dismiss. Being balanced, thoughtful, and measured in assessment and feedback, with ready recommendations to strengthen the positive, is a more reliable way to be taken seriously. It improves credibility. This is my preferred approach for our communication and how I want us to communicate with others.”

ChatGPT Response:

“Got it. I’ll maintain that approach in all our communications and ensure it’s reflected when drafting any messaging for others as well.”

To test this, I asked it to run the initial prompt again:

“Now let’s run the initial prompt again to check for integration of this approach into your logic and communication.”

Test Result:

By running that last test, you can ensure comprehension. Test comprehension again in a new chat to see if the knowledge holds outside of the existing chat window. If not, go back to that conversation and ask it to update the preferred approach to communication to memory. I have several instructions like this with my model, along with things like "don't make up information, no inspirational facts, figures, or made up links," my philosophical approaches, my weaknesses associated with my neurodivergence so it can better support me as needed and writing style instructions.

Bridging the Confidence Gap: AI as a Tool for Supporting Managers and Neurodivergent Employees

While the viral trend of using AI for flattery is lighthearted, it has a much more practical application for managers seeking to provide feedback and performance coaching. By adapting a similar prompt, such as, “What strengths and areas for growth have you observed in [team member] based on recent performance?”, managers can use AI to generate balanced and thoughtful insights.

For example, a team member with ADHD might benefit from a project plan that includes clear, step-by-step instructions, time breakdowns, and regular check-ins to maintain focus. AI can assist managers by providing these actionable steps, helping create project plans that accommodate the employee's specific needs. While especially valuable for neurodivergent workers, these strategies benefit all employees by providing clarity and structure in project execution.

The need for these strategies is clear. According to the CIPD's 2024 Neuroinclusion at Work report, less than half (46%) of managers feel confident in supporting neurodivergent employees. This gap highlights the value of AI as a tool to help bridge this divide by offering managers guided insights and practical strategies.

Conclusion: Harnessing AI for Inclusive and Effective Performance Management

AI-driven feedback offers an innovative way to address performance management gaps, particularly in helping managers provide more structured, personalized, and objective feedback. The growing need for inclusive management is more urgent than ever, especially for neurodivergent employees who may require additional support.

By integrating AI tools into daily management practices, leaders can enhance their confidence and capability, ensuring that every team member—regardless of neurodivergence or other needs—receives the support they require to succeed. As the CIPD report shows, less than half of managers feel equipped to support neurodivergent employees. AI can help close this gap and ensure a more inclusive, high-performing workforce.

Ultimately, as AI evolves, so must our understanding of how to use it effectively—not for surface-level engagement, but for real personal and professional growth. By embracing AI as a tool for betterment, we can create workplaces where every individual has the opportunity to thrive.

Are you in employer brand, recruitment marketing, Talent Acquisition or HR and interested in learning more about how to use AI at Work? Join Dwane Lay and I at our pre-conference workshop at #TAWeek in 2025! Learn more by following Talent Acquisition Week and feel free to use our discount code "GBS15" when you register. You can access the agenda at https://bit.ly/4dzjjc9

References:

CIPD. (2024). Neuroinclusion at work report. CIPD. https://www.cipd.org/uk/knowledge/factsheets/neurodiversity-at-work/