Director, Learning Design & Development| PMIAsheville, NC, United States
Validating and checking outputs is critical when working with AI systems like Generative AI. Such validation approaches may include establishing clear criteria, implementing strong testing protocols, and continuous refinement.
In your experience with AI, what are some best practices for ensuring the results you receive are accurate, relevant, and aligned with your original goals?
Rutwij ChaukhandeDelivery & Strategic Execution Office Lead| State StreetHyderabad, Tg, India
Contextual information, Specificity, Understanding of the persona and Command on the underlying data fed into the AI model. Saving Changes...
Mark CrudoDirector Professional Services Strategy & Transformation| ICU MedicalChicago, IL, United States
Ensure that your goal or solution is well-defined with well-defined/logical details and supporting information/data reviewed and prepared by appropriate stakeholders. Some level of checklist or traceability confirming the AI responses back to the original request’s requirements should be confirmed to prevent AI false assumptions/inaccurate results. Be prepared to iterate and adjust to align with human factors or other contributing circumstances. Saving Changes...
The prompt formula plays a role here, the context and simplicity of the prompts should be considered.
Review the results for accuracy and find/prompt for sources of the results provided. Saving Changes...
Using a prompt formula to clearly articulate the task, checking the output and then refining where necessary to ensure accuracy and alignment. Saving Changes...
In every interactive communication either with humans or machines, the quality of response or result from interaction is synonymous to the inputs or promptings; just as the saying "garbage in garbage out". so to achieved the best result or output, a strategic plan for a concise, clear, contextual promptings, format and expected outcomes refined continually should yield the desired outcome. Saving Changes...
Lakshmy ShankarPM I| IBM INDIA PVT LTDMumbai, Maharashtra, India
I would consider some best practices as below for ensuring the results you receive are accurate, relevant, and aligned with your original goals..
1. Define the persona or character for which the question is being asked
2. Define the request in simple and precise terms providing a context
3. Provide some examples if possible
4. Define the type of output that is expected from the AI tool
5. Validate the output provided to check if it meets the original goal. Saving Changes...
Some of best practices for ensuring the results you receive are accurate, relevant, and aligned with your original goals when using AI systems are:
1. Be aware of the limitations of the AI system you are using. Understand what it can and cannot do, and don’t over-rely on it for tasks outside its capabilities.
2. Maintain a level of human oversight, especially for critical tasks. AI should augment human decision-making, not replace it entirely.
3. Regularly track the performance of the AI system against your objectives. Adjust your parameters, prompts, or data sources as needed to maintain alignment with your goals. Saving Changes...
To ensure accurate, relevant, and aligned AI results as a project manager:
Define Clear Objectives: Clearly articulate goals and set expectations.
Choose the Right AI Tool: Select an AI system that matches your needs and understands its purpose.
Use Quality Data: Provide clean, relevant input data with proper context.
Test and Validate: Run pilot tests, compare outputs, and review results regularly.
Incorporate Feedback: Use iterative feedback from stakeholders and AI outputs to refine processes.
Ensure Transparency: Document the AI process and ensure decisions are explainable.
Monitor for Bias: Actively detect and mitigate biases, considering ethical implications.
Plan for Contingencies: Have backup solutions if AI outputs are unreliable.
Stay Informed: Continuously learn and adapt to new AI developments and best practices. Saving Changes...
I agree with all the responses above, i will just add that you continuously monitor your input and check the output against multiple latest information. Saving Changes...
Ask the AI, what are the question relevant to refine the prompt that are missing or required to obtain the proper context required to the Goal or task requested. AI works better with hegel or socrates, so, with the proper questions answered well, it will elevate the temperature and accurace of the request. Saving Changes...
"In the real world, the right thing never happens in the right place and the right time. It is the job of journalists and historians to make it appear that it has."
PMI Team Member
