Generative AI Literacy

Does company policy allow AI use?

Check Company Policies! Some companies are actively using GenAI while others may have specific policies in place, or outright bans on the use of the technology in order to protect data, customer information and even trade secrets. If you don't know the policy, check with your supervisor, colleagues or human resources before using GenAI in your work.

If no, Don't use AI

If yes:

Are there specific tools you must use?

What tools should you use? If you get the go ahead to use GenAI, be sure to ask if there are preferred tools. Your company may have paid access to a tool, be using tools customized to protect privacy, data and trade secrets, or have a preference or process/procedure for using and acknowledging particular tools. Know before you start!

If there are no specific tools required, Consider data and privacy when selecting tools! If your company doesn't have a preferred tool, but is open to GenAI use, now's the time to demonstrate your skills. Review prospective tools and how you might use them. Be sure that the selected tool will actually provide benefits, and ensure that the information you'll need to submit to the tool (prompts, background info, personal info for sign up, etc) do not violate any privacy, data protection or trade secret concerns before using. Don't use GenAI without considering these issues.

If there are specific tools you must use, or once you've reviewed privacy and data concerns:

What tools should you use? If you get the go ahead to use GenAI, be sure to ask if there are preferred tools. Your company may have paid access to a tool, be using tools customized to protect privacy, data and trade secrets, or have a preference or process/procedure for using and acknowledging particular tools. Know before you start! Then...

Have you checked that the output is accurate?

Evaluate the Output! Just as you would if you were permitted to use GenAI for an assignment at College, make sure you check and evaluate all generated content to ensure it's appropriate for use at work. You'll want to verify any sources, fact check, and watch for fabricated details or misinformation. Taking a few moments to do this can boost your reputation!

If No: don't use AI.

If yes:

Have you cited or disclosed AI in your work?

Do you need to cite or disclose the use of AI? Before you finalize any work with GenAI assistance, make sure you understand your employer's expectations. Do you need to cite the GenAI tool? Is there a standard way of disclosing the use of these tools? How would your employer like this handled? Make sure you can put the finishing touches on your work with proper acknowledgement!

If yes... AI use is possible!

• social media content
• planning and building marketing strategies
• search engine optimization (SEO)
• content ideation (brainstorming, etc.)
• writing content
• content research
• proofreading
• bug-fixing and debugging software
• testing code snippets
• code generation and research
• drafting messages to customers
• analyzing customer feedback

Those in healthcare may think this is all about writing and coding, but AI is transforming the healthcare industry as well. In addition to the above, AI models are now:

• automating documentation;
• helping with data entry and extraction;
• managing communication;
• monitoring regulatory compliance;
• helping with administrative workflows and task prioritization;
• facilitating patient outreach;
• image enhancement for better diagnosis;
• data augmentation;
• noise reduction and pathology prediction;
• personalized treatment plans;
• clinical support;
• research and development of new drugs.

Even more hands-on patient help, such as the tasks normally undertaken by nurses and other practitioners, is becoming transformed by GenAI, including:

• patient data analysis;
• predicting potential complications;
• real-time suggestions for interventions and medication dosages;
• documentation assistance, including clinical notes from voice recordings and summarizing patient interactions;
• simulating patient scenarios for training;
• patient communication, including real-time translation services;
• generating personalized patient education materials.

Those in the automotive repair industry will begin seeing GenAI as well:

• analyze vehicle data, symptoms, and repair history to suggest potential issues and solutions more accurately;
• interpret complex diagnostic codes and sensor readings;
• analyze images or descriptions of parts to identify them accurately;
• anticipate parts needs based on common repair patterns, improving inventory management;
• create realistic simulations and training scenarios to practice complex repairs;
• help mechanics explain technical issues to customers in simpler terms;
• provide estimated repair times and costs more accurately;
• analyze repair shop data to optimize workflow.

Completing thorough audience and context analyses before beginning will help you to narrow the focus of your work, address the needs of the audience, and adapt your communication strategy to the situation at hand.

Clearly articulate your goal, and reflect on how the genAI model can assist you in achieving it. For example, your objective may be to write a proposal to persuade a client to accept a plan to develop and provide genAI training at the organization. Ask yourself: Are you responding to a request, analyzing data, addressing an opportunity or problem? Are you creating a proposal, sending correspondence, creating a presentation or video?  It might help at this stage to think of achieving your goal in a step by step manner in order to plan an efficient research, drafting, and revision strategy and determine how the model can assist at each stage.

Tell the genAI tool what you expect it to produce: “Draft a an outline for a proposal for the development and provision of genAI training at a mid-sized financial services company.” Ask yourself: Are you requesting a document outline, a summary, a draft video or presentation, ideas for a topic, or a chart or graph that illustrates a point?

Collect the information and data you will be using to create the prompts. You will use whatever information you have available to inform and guide the genAI model in focused tasks. Microsoft Copilot, when widely available in enterprise systems, will greatly facilitate this process by automatically accessing your documents, emails, meeting notes, and other materials as it processes requests. As of this writing, in most workplaces, this still must be done manually.

Choose the genAI model that will best achieve your goal by considering the following factors: degree of reliability, cost, in-application tools, accessibility, and privacy policy. Many organizations will provide you with approved tools, so your choice of models will be limited to what is available. At Seneca, Copilot (connected to ChatGPT 4 and the internet) is available through Seneca accounts. In other contexts, you have great latitude in the choice of tools to use: open source or commercial, by subscription or free, by version, and you can choose based on capacity, features, ethical practices, capability, and the like.

 

Review your organization’s privacy, confidentiality, and other applicable policies, so you can work within its legal and ethical standards when using genAI applications and their output.

Review prompting techniques. Does the task at hand lend to conversational or a more complex structured prompting? Choose the method that will best help to achieve your goal. Doing so will reduce the amount of time and the number of iterations required.

You may want to try inputting a prompt that encompasses the entire task, or you may choose a phased strategy that involves breaking up the task into sequential stages. Again, think of what you want to achieve and decide on the best approach.

The more specific the prompts, the more likely the genAI model will provide outputs that match your needs. Use your planning data and information to create prompts that contain information about context, problem or need, criteria, goal, etc. The more information you provide regarding context and topic, the more useful the output.

Review the outputs and refine your prompts to achieve more precise, goal-oriented responses. The first output is usually not considered the best, so engage with the application in a conversational manner and ask it to refocus its work using additional prompting. If the model does not provide you with the necessary output, consider using a different one.