Embracing the Human Factor in AI: Insights from Leigh Coney

The Paradox of Sycophantic AI

AI models are notoriously viewed as "yes-men." By design, they tend to agree with us, echo our thoughts, and support our ideas. However, this inherent sycophancy is a double-edged sword. While it can provide validation and support, it often leads to a lack of critical thinking. As Leigh Coney, a 34-year-old AI consultant based in France, points out, one of the major pitfalls in many AI projects is the overlooking of the human factor. This oversight results in a replication of human biases, ultimately stifling creativity and clear thinking.

Transitioning from Psychology to AI

Coney’s own journey into the world of AI began when Microsoft announced Copilot for its suite of products. Recognizing the potential for AI to permeate every business sector, she transitioned from teaching psychology to becoming an AI consultant. Her background in psychology allows her to tap into human behavior, ensuring that AI interactions are more nuanced and productive. As she builds custom AI automations and agents for diverse industries, she integrates psychological principles to enhance efficiency and promote growth.

Challenging Your Ideas with AI

Coney emphasizes the necessity of asking AI to challenge our ideas. A typical request made to a model like ChatGPT may not unearth the flaws in a plan. Instead, Coney asks AI specific prompts, focusing on potential assumptions and inviting scrutiny. Questions like “What am I overlooking?” or “Where am I not being clear in my thinking?” provoke deeper insights, helping users refine their strategies.

Tailoring AI Responses to Your Audience

Another strategy Coney employs is specifying the audience when using AI. For instance, if you’re preparing to pitch an idea to a CFO, telling the AI to “Act as a skeptical CFO and ask five challenging questions” prepares you for a tough critique. This approach not only arms you with better insights for your pitch but also forces you to confront possible weaknesses in your proposal before facing actual scrutiny.

Utilizing the Framing Effect

The framing effect—how information is presented—can significantly influence perceptions and decisions. Coney illustrates this with the example of crafting messages. When delivering bad news, the wording can drastically affect the message’s reception. Rephrasing a project update to emphasize resilience and learning can inspire a more positive outlook than a straightforward acknowledgment of setbacks. Coney finds that tweaking her prompts in AI chats—sometimes altering just a word—can yield vastly different and often more constructive responses.

Evolving Attitudes Toward AI

Initially, Coney harbored concerns about AI’s rapid advancement and its potential impact on jobs. However, a recent underwhelming release of a highly anticipated model, such as GPT-5, led her to reassess. She believes there’s still ample time for the workforce to adapt, suggesting a calmer, more measured approach to the evolving landscape of AI. This shift indicates the need for society to focus on upskilling and enhancing human-computer collaboration rather than fearing obsolescence.

The Importance of Understanding Cognitive Biases

For Coney, a deeper understanding of cognitive biases—those quirks in our thinking that can distort our decisions—is invaluable. This knowledge enhances not only human communication but also how we interact with AI models. By acknowledging our biases, we can sharpen our thinking and significantly improve the output yielded from AI systems.

Through thoughtful application and careful interaction with AI, we can harness its capabilities while mitigating its shortcomings. Coney’s insights remind us that the bridge between human intuition and artificial processing can lead to breakthroughs when approached with mindfulness and critical engagement.

---

If you’re an AI expert with insights to share, Leigh urges you to reach out, reinforcing the communal nature of knowledge in this rapidly changing field.