When using artificial intelligence to assist in any tasks, it’s important to approach the technology with the right mindset. AI is bad at handling 10 tasks simultaneously. But it's very good at doing one thing at a time.
This tutorial will explore a fundamental principle in harnessing AI effectively: giving AI tasks, not jobs.
This approach will help you optimize your interactions with AI tools and give you the best possible results.
Understanding Jobs vs. Tasks
Let's begin by distinguishing between a "job" and a "task." A job can be seen as a complex objective that requires multiple steps to complete.
For instance, creating an entire lesson or an end-of-unit test for a course constitutes a job. On the other hand, a task is a single, well-defined action within that job. AI tools excel at performing specific tasks rather than taking on entire jobs.
Example
The key to success with AI is to break down a comprehensive job into smaller, manageable tasks.
Let's take the example of creating an article, something I personally do very frequently with the help of AI.
1 - You're a tech writer. You're going to write an article about OpenAI's new model. Give me the main parts of the article (from 3 to 5), without elaborating on their content. Use this data to help you make the parts: [ADD CONTENT] (submit)
2 - For each part, give me 3 examples of the part with its subparts, I'll choose the one I prefer. (submit)
3 - For each subpart, make a factual first sentence in bold, then two factual sentences that complete the first. 3 sentences. Speak in the past tense. If you use notions of time, be precise (mention the date). (submit)
4 - Make sure you don't talk about the same event twice. Don't discuss, examine, or evaluate anything. Each sentence has to be factual. (submit)
The key here is to ask the AI to perform each task once it has successfully answered the previous question. This will also enable you to check that the answer is what you expected.
Tips for task-based prompts
Here are some tips to enhance your interaction with AI when assigning tasks:
- Be Specific: The more detailed and specific your prompts, the better the AI can perform the task. For instance, instead of asking for "multiple-choice questions," specify "10 multiple-choice questions on Newton's first law of motion."
- Iterate and Select: AI might not deliver perfect results on the first try. Be prepared to iterate, refine the prompts based on the initial outputs, and select the best results.
- Combine Outputs Thoughtfully: Once you have the AI-generated components, it’s your role to integrate them thoughtfully into a coherent whole. This may involve editing for consistency, relevance, and impact. That's why a human is always needed to leverage the power of AI.
Conclusion
AI is a powerful tool, but it requires a nuanced approach to be most effective.
By understanding the distinction between jobs and tasks, you can leverage AI to perform specific, well-defined tasks, thereby enhancing the quality and relevance of the content they produce. This approach not only saves time but also integrates AI seamlessly into any workflow.