You may have heard that artificial intelligence is going to 1) take all the jobs, 2) run local governments into the ground, or 3) may possibly end all human existence. As an information professional, I believe these fears are overblown and unfounded, especially with the types of AI that are available for public use currently. At the moment, AI, especially generative AI, which are computer science-based models that have been trained on systems that vaguely resemble the connections in human brains, is able to do many things that are all based on ONE vital ingredient – prompts. Prompts are the foundation of everything AI can now do, like create written text from written or spoken inputs, make audio tracks, images, and short videos, and other rapid, repeatable, and surprisingly accurate content creation functions. These models are not intelligent the way humans are (there is a lot of debate about how AI does what it does at all)–their skill is in mimicking one function of human output and, in many cases, repeating its creation with more speed than a human, when they work correctly. Humans that can interface with AI can increase their productivity.  

There is a rule of thumb being proposed to avert any sort of wide-spread negative consequences from generative AI; that is the concept of “keep a human in the loop”. This means making sure a human is included in sign-offs, final says, results-checking, or other functions that can prevent generative AI tools from acting without a human authorizing its actions or steering the AI tool back on course if it has gone astray (e.g. producing results increasingly dissimilar from those that its operator desires). So how does a human operate and interact with generative AI tools? 

Before we get into that, there are a few other rules that generative AI users should always remember: never enter personally identifiable information into generative AI tools. This applies to your name, social security number, address, phone number, but also could be the school you are attending, your car’s license plate number, or your annual taxable income from a previous year. These facts may clue someone into your identity, your location, or your passwords. The second rule is: do not expect anything that an AI tool generates to be truthful. If you are asking questions about medical issues, finances, legal proceedings, or other processes that involve important decisions, you will want to have other sources of trustworthy information with which to verify what the AI tells you. The librarians of the Northern Onondaga Public Library are happy to help patrons find trustworthy information sources that can be used to verify AI-generated information.  

This leads to another important consideration: since audio, video, images, and text can all be rapidly created with little effort on the user’s part and easily tweaked to produce a desired response, it is extremely important to interrogate the validity of any sound, image, video, or text that produces a strong positive or negative reaction in you, especially if the reaction is emotional. These tools are being used to manipulate people. You must be vigilant as humanity has advanced beyond the assumption that video, images, and audio are always sources of truth. 

One important note for all users of AI tools but especially for artists, musicians, and all content creators that share their creative works on the internet: in order to create AI tools, content shared on the internet and created while using AI tools is recycled (sometimes contrary to the creator’s intentions) into the billions of works being used to train the largest AI models. Some services allow users to opt out of having their creations be used to train AI, however, many services do not have that option. Beware that your interactions and uploads are being watched and logged by AI models. To do a non-comprehensive search to see if your information has been used to train AI, try the “Have I been Trained?” tool. For more information about opting out of your data being used to train AI, please visit this helpful page: 

https://www.linkedin.com/pulse/your-comprehensive-guide-ai-opt-outs-paul-j-ashton-pcp0c/  

Generative AI tools are here and they are just that, tools. Just like a sharp knife is better for chopping, it can be dangerous as well. AI has done a lot of great things, including unlocking protein folding and cancer screening for better medical outcomes; assisting people, including individuals with disabilities, with everyday tasks; monitoring roads and water supply infrastructure to prioritize repairs and fix hazardous conditions. It’s also been applied to scientific uses such as wildlife conservation and categorizing astronomical data to spot near-earth objects and much more– even identifying if a work is the product of AI!  

Generative AI tools can be used to communicate these important advances, garner more funding for these great endeavors, and grow staffing and support for breakthroughs that will save lives, and improve quality of life and longevity. It can also be used for great harm. You get to choose what you do with AI tools, but ultimately AI is only as good or bad as the humans trusting, wielding, or exploiting it.  

We can evaluate AI tools before we use them.

Open Library provides such a framework with an AI evaluation rubric that has been turned into a fundamental quiz. There is also an in-depth quiz further down on the same page if you want to rigorously evaluate your generative AI tool to the highest standards. This evaluation is meant for teachers, so there are some questions about usefulness of the tools for teaching. You can choose “minor concerns” if those questions do not pertain to your application.  

The ethics of using some AI models are currently in question. There have been news stories about exploitation, unethical firings, copyright infringement, environmental exploitation, and other ethically questionable practices by companies and subsidiaries in charge of many major generative AI models. AI tools can also exhibit bias. NOPL advocates the responsible use of generative AI tools and does not endorse any specific model. There are models that are more transparent about how they work and provide more privacy control, and there are models that provide specific benefits and perks for users. We cannot suggest one model that will be best for everyone’s needs. We will highlight the models that are easiest for beginners.  
 

Getting Started: 

We recommend interactive programs from GetSetUp to learn the basics. Click here to view upcoming offerings about AI from GetSetUp. LinkedIn Learning, which you can access for free with your library account, also provides a great overview of generative AI in the course: “How to Research and Write using Generative AI Tools.”  

Follow the steps below if you feel like diving straight in to using generative AI tools. 

ChatGPT – No smartphone or computer is required to access this generative AI interface. The simplest way to use it is to call 1-800-242-8478 (ChatGPT). This interface is speech to text input. Your phone call is transferred to text and the ChatGPT interface is asked for a text response which is turned into voice to be played over the phone line. If you do happen to have a smartphone, these conversations can be recorded and transcribed for later reference.  

  1. Dial 1-800-242-8478 
  2. Wait for the tool’s voice to ask for your question or query. 
  3. Ask: “Please list 7 activities to do with 10-year-olds that require no money” (or whatever question or query you may have)  
  4. Listen for the response.  
  5. Ask a follow-up question or hang up. 
  6. Repeat steps 4 and 5. 

All of the preceding steps can be done via text on a computer or smartphone with an internet connection at chat.com. You can ask the tool whatever question you like, keeping in mind the important rules listed above. Some generative AI models are programmed not to provide answers for topics that its creators think are incendiary or otherwise contentious or harmful. You may run into these barriers. Try reframing or rephrasing your questions.  

For more information about creating more precise queries to receive more exact responses, check out this video of a recent NOPL program. More advanced techniques are discussed in this LinkedIn Learning course: “Introduction to Prompt Engineering for Generative AI” that you can access online. 

 

 

There are several tools available for image creation. The easiest is probably Google Gemini. 

  1. On a computer or smartphone with an internet connection, go to: https://gemini.google.com/  
  2. Locate the box where you can ask your question or query. 
  3. Type into that box: “Please generate an image of a penguin winning the lottery” (or whatever query you may have) and type enter. 
  4. Wait for the generated image to load.  
  5. Add follow-up changes or details or stop using the tool. 
  6. Repeat steps 4 and 5. 

 

 

There are also several tools available for audio creation.  ElevenLabs is fairly easy to use. 

  1. On a computer or smartphone with an internet connection, go to: https://elevenlabs.io/text-to-speech  
  2. Locate the box where you can enter the text to create into speech. 
  3. Type into that box this excerpt from Shakespeare’s Sonnet 104:
    • “To me, fair friend, you never can be old,

      For as you were when first your eye I eyed,

      Such seems your beauty still. Three winters cold

      Have from the forests shook three summers’ pride,

      Three beauteous springs to yellow autumn turned

      In process of the seasons have I seen,

      Three April perfumes in three hot Junes burned,

      Since first I saw you fresh, which yet are green.”

      (or whatever query you may have)  

  4. Wait for the generated audio to load.  
  5. Add follow-up changes or details or stop using the tool. 
  6. Repeat steps 4 and 5. 

 

 

Create video content: Sora or https://www.moonvalley.com/ (both forthcoming in the US) 

 

 

Here are some other helpful AI tools. This list demonstrates the range of processes that a user can perform with AI aid, and there are hundreds perhaps even thousands of other tools available. 

Scale up images: https://upscayl.org/  

Brainstorm (many models can do this): https://www.perplexity.ai/  

Create process documents: https://scribehow.com/  

Simplify complex topics or make a to-do list: https://goblin.tools/  

Create a recipe from on-hand ingredients: https://goblin.tools/Chef  

Detect if something was created with AI (can be inconsistently reliable and biased):

https://www.truemedia.org/

https://gptzero.me/

https://www.grammarly.com/ai-detector

https://libguides.marian.edu/c.php?g=1321167&p=9721351

 

Here is a handy list of new titles about AI that are available at the NOPL libraries.