Main image of article Using Soft Skills to Boost Your Generative AI Usage

Summary

You probably have some questions about how generative AI platforms such as ChatGPT will end up impacting your job. There’s also a lot of FUD surrounding artificial intelligence, including the widespread belief that the technology will kick off massive layoffs in the next year or so. While that almost certainly won’t happen, it’s worth examining how you can effectively collaborate with these tools as they find their way into tech stacks everywhere.

Because so much of generative AI work is “prompt engineering,” which means asking questions and expecting ChatGPT or other platforms to generate a useful response, your “soft skills” such as communication can come into play in a big way. Let’s break it down: 

Generative AI platforms are powerful tools, but they’re not Skynet; they’re not a super-intelligence that knows instantly what you want. If you want to leverage your communication skills for better prompt outcomes, keep this in mind:

  • Specificity is Key: Don't expect AI to decipher vague prompts. Be clear and concise about your desired outcome. For code generation, provide detailed descriptions of functionalities and desired outputs. For QA, outline specific scenarios and edge cases you want tested.
  • Embrace Structured Language: Many AI platforms leverage natural language processing (NLP) but excel with structured formats. Utilize code comments and clear test case descriptions to enhance communication with the AI.
  • Provide Context: AI thrives on information. For code generation, share existing code snippets, APIs, or relevant documentation. For QA, outline the project's purpose, target audience, and any specific industry regulations.

Again, because generative AI isn’t the ultra-intelligent AI we’ve seen in countless movies, it doesn’t intuitively understand what exactly you want. Even though it’s not human, you can still use your empathy to get the most out of your particular chatbot platform:

  • Recognize Bias: AI models are trained on vast datasets, which can inherit societal biases. Be aware of these potential biases and review outputs with a critical eye.
  • Embrace Iteration: AI outputs are rarely perfect on the first try. Approach AI as a collaborative partner, refining prompts and parameters based on initial results.
  • Focus on Intent Over Perfection: Sometimes AI-generated code or test cases might not be aesthetically perfect. Focus on whether the code functions as intended or the tests achieve their goals.

With generative AI, many people take a ready-fire-aim approach… but your results can be so much better if you carefully consider your step-by-step strategy before engineering your prompt. Let’s look at two scenarios:
 

  • Scenario: Code Generation: You need to generate code for a login functionality.
    Effective Approach: Clearly define the user types (admin, user), login credentials (username/password, social media login), and desired functionalities (session management, error handling). Run the AI with this information and review the generated code. Be prepared to iterate on the prompt if needed.
     
  • Scenario: QA Testing: You want to test a mobile app's login functionality on different devices.
    Effective Approach: Explain the app's purpose and login process to the AI. Provide information on various device types (Android, iOS, different screen sizes). The AI can then generate test cases that simulate login attempts across those devices.

AI is a powerful tool, but it lacks real-world context; it doesn’t actually understand what you want or need. It’s up to you (and your critical thinking and problem-solving skills) to show it what you expect. But if you can master the burgeoning art of human-AI collaboration, you can potentially make your workflow more efficient than ever before.