5 Principles Everyone Should Know When Applying AI Tools to UX Design.
#1 Create Specific Design Requirements
Prep before using the chosen A.I tool. Spend time on developing design requirements. Design requirements act as a parameter. The more specific these are the more you will be able to gauge whether or not the output fits within those parameters. If you don’t have any parameters or a specific goal that is going to impact the accuracy of the prompt you give it thus impacting if the output hits the target you’re looking for.
#2 Prompt Accuracy
Write prompts that precisely reflect what you are looking for. Here are some ways you can write on the ball prompts.
Here are some ways to write good prompts…
Be clear and specific.
Clearly state your request. Avoid being ambiguous because it can lead to unclear or just off target outputs. A good way to do this is to specify the TYPE of information you are looking for.
Provide Context
Include relevant details in the prompt you input. This will support the tools understanding of the task at hand. Imagine you are explaining to a tutor what you need help with. The AI tool cant read your mind so therefore you must give as much detail as possible or it wont be included in the output.
Use Complete Sentences
using complete sentences can help ensure that the output is also coherent. So avoid using fragments or incomplete sentences.
Use Specific Nouns
Avoid pronoun ambiguity. Meaning, when you are referring to a specific noun in a prompt it is beneficial to use the actual name of the thing instead of pronouns. This is especially important when the noun that is being replaced with a pronoun is not mentioned in the prompt. You are basically minimizing the risk in misunderstanding the prompt and missing the target.
Ask One Question at a time
If the query is too complex then chunk it out. Break the complex query or task into individual pieces.
#3 Regenerate
A.I tools often have a “regenerate” or a “try again” option that allows you to generate alternative outputs. If the first one doesn’t float your boat dont hesitate to use this feature. Resist the temptation to accept subpar outputs. Instead of settling for the first one use the regenerate option to explore different outputs until you reach one that meets your expectations. If a output completely misses the mark and fails to meet your crieteria reject it. A.I tools are designed to assist and its actually beneficial to communicate your expectations clearly by accepting only outputs that align with your goals. Most importantly DO NOT compromise your vision. AI is a tool not a decision maker…thats you. Maintain a steadfast commitment to your creative problem solving vision that uses AI as a helpful assistant in achieveing that vision.
#4 Make Changes to The Output
The output is a starting point not a finished product. Don’t ever expect AI to give you a finished product almost every time you use it to generate a design deliverable you will have to make thoughtful changes to the output. The exception being if the tool is used to create an output that serves as a foundational element or just a fun experiment. AI generated outputs should be viewed as a starting point rather than a conclusive solution. Design is iterative in its nature and the process relies on refinement and adaptation based on insights gained during the process.
#5 Be Aware of Bias
always keep an ethical lens on while utilizing AI.Even though Gen AI creates new and origins content, Keep in mind that everything it’s going to output is a reflection of the data it was trained on. There are some really easy ways to mitigate these biases that are bound to occur.
Understand AI limitations
While using any AI tool keep in mind that the tool will have its limitations. Work on understanding those limitations and take note of the potential biases that occurs while using it. This awareness will help you interpret and critically evaluate the tools’ outputs.
Question AI outputs and recommendations
Whatever you do don’t just blindly accept the generated outputs and recommendations. Like all information, it’s important to question and validate the suggestions provided with critical thinking.
Check for diverse outputs
Test the tool with different inputs to check that it has the capability of producing a diverse set of outputs. This can help identify potential biases that may arise in specific contexts.
Document Design Decisions
document all key design decisions but absolutely document and note decisions based on AI recommendations. If you’re working with a team this can help give light onto the rationale of the decision. Also if there is no validating rationale as to why the decision was made except that AI generated it then that’s a red flag to watch out for.
