AI specialist discloses five keys for composing optimal ChatGPT and similar systems' instructions
Meet Thiago Goldschmidt, a tech whiz who is the Head of Business Development at Visio Lab. His work is practically a dance with AI – he uses AI in almost every aspect of his tasks. Apart from his role at Visio Lab - a company that develops an AI-powered self-checkout system for the food industry - Goldschmidt also guides other companies on the AI front.
For him, AI is like a super-smart 24/7 mentor if you know how to employ it right. What AI generates depends on the question asked. Many people input a question and are amazed by the output, but according to Goldschmidt, the real relationship with AI begins after the response.
He calls this collaboration 'prompting.' For Goldschmidt, crafting a good question is a collaboration with AI.
Goldschmidt's optimal question comprises of five elements: RAKBI.
Role: Define who the AI should act like. Goldschmidt advises visualizing the AI as a specific person or professional to get better results.
Task: Describe what the AI should do. This includes setting the objective, desired output, and specifying the context.
Context: Provide background information about the topic, problem, data, or scenario the AI is dealing with. This helps the AI understand the nuances of the question.
Evaluation: Define the criteria for assessing the output. This helps guide the AI towards the type of response that's most useful or relevant.
Iteration: Be ready to fine-tune your question based on the initial response. Tweaking the question can lead to better results over time.
Remember, AI is a tool, and like any tool, its output depends on how it's used. So, get ready to become an AI whisperer with Goldschmidt's RAKBI tips!
Technology and artificial-intelligence play a significant role in Goldschmidt's work, as he uses them to achieve his tasks efficiently. If you want to reap the benefits of AI, Goldschmidt suggests approaching it as a collaboration, where you craft a well-structured question using his RAKBI method – Role, Task, Context, Evaluation, and Iteration – to elicit accurate and useful responses from the AI.
