Utilizing GenAI tools effectively involves choosing the right tool for the task at hand, crafting prompts that generate the output you want/expect, AND using the tools ethically and with integrity. While this section of the guide focuses on crafting prompts and evaluating and choosing specific tools, be sure to fact-check any AI-generated content and sources you plan to use in the work you share with others and/or publish. Students should only use these tools within the guidelines established by instructors and the Academic Integrity Office. Employees should follow guidelines established by their employer, whether that is UC San Diego or another entity. Make sure to document/cite the parts of the work created by the GenAI tool. Do not include anyone's Personal Identifiable Information (PII) in your prompts, whether it is your own or someone else's.
Learn Prompting defines "prompt engineering" as "the process of communicating effectively with an AI to achieve desired results."
Using GenAI tools effectively requires that the user know the right questions to ask, and how to phrase them for the best results. Vague or generic questions generate vague or generic results. (In other words, garbage in, garbage out.)
Tips for crafting prompts to get the best results from chatbots:
Many tech reviewers have published comparisons between popular AI chatbots, but the landscape changes rapidly.
This guide primarily features free and freemium tools. We recommend experimenting with these options before investing in any paid subscriptions. Unless you need a feature that is unique to another tool, we also highly recommend choosing chatbots that connect to the internet and provide links so that you can verify their claims.
Meta's popular, open source Llama model powers many different tools. Meta AI's assistant is available as a stand-alone app and infused throughout Meta's products like Facebook and Instagram. The assistant in Facebook, Instagram, and WhatsApp messengers is designed to provide personalized and empathetic responses, understand context and nuances of human conversations, and offer creative solutions and ideas.
If you are working on a project that requires citations, you want to confirm the generative AI chatbot's information sources, or you are seeking information newer than would be included in the chatbot's large language model (LLM), you will want to use a generative AI tool with internet search capabilities like Microsoft Copilot (Bing Chat) or Google SGE (Search Generative Experience).
Dimensions is a linked research data platform, including information on publications, research funding, patents and clinical trials - all on an international scale. Dimensions is designed to help anyone involved in research to gain better knowledge about activity in their research area or institution, and how this is changing over time. Includes AI summarizer feature, Dimensions Research GPT (Dimensions’ scientific evidence + ChatGPT), and Chat with this PDF in Papers integration.
Some examples of the ways in which Dimensions can be used include -
• Discovery of the latest publications, awarded grant funding, clinical trials or patents on any topic of interest, worldwide.
• Benchmarking of research organizations, funders, publications or researchers in a particular topic, or across all research activity.
• Identification of new sources of research funding for future funding applications
• Clear and unbiased analysis of the research activity at our own organization, and how this is changing over time.
Other AI-powered tools to improve your writing. Students take note: please make sure that you are using these in accordance with your instructor's rules on use of generative AI and the Academic Integrity Office's guidelines. If you have questions or concerns, please consult your instructor first.
Most AI image generating sites operate on a freemium model where users are allotted a set number of credits. For many tools, the credits reset daily or monthly, but some tools provide only a one-time allotment. Users who pay for accounts can generally unlock more credits and/or premium features. We do not recommend paying for any sites without further research.
DALL-E is OpenAI's text-to-image model. DALL-E 3 is included in a ChatGPT Plus subscription or available for free via Copilot Design.
FLUX.1 is a suite of advanced text-to-image models known for their exceptional image detail, prompt adherence, style diversity, and scene complexity. FLUX is developed by Black Forest Labs, which was founded by former researchers from Stability AI.
Stable Diffusion is a family of popular open source text-to-image models created by Stability AI. Anyone can download and run the code on their own PC, and a plethora of generative AI sites have incorporated Stable Diffusion code into their tools in both its original form and modified through low-ranking adaptation (LoRA), an AI training technique that helps to fine-tune the models.