AI and the Library. Week Four. Perplexity AI – Tool Review.

You probably don’t need us to tell you—there are a lot of AI tools out there. While ChatGPT might have grabbed the headlines early on, Perplexity AI is quickly becoming one of our favourites here in the library.

Perplexity AI is making waves in the world of information retrieval, and for good reason. Unlike its more conversational counterpart, ChatGPT, which often indulges in lengthy dialogues, Perplexity adopts a direct and efficient approach to answering queries. This makes it an ideal tool for those who need quick facts, thorough research, or insights into specific topics. Not only that, but it provides citations for all its responses.

Why Choose Perplexity?

Here are the standout features that make Perplexity a top choice:

  • Real-Time Information: Perplexity pulls fresh data directly from the web, ensuring you always get the most current information.
  • Clear Summaries: Instead of bombarding you with endless links, it delivers concise, direct answers to save you time.
  • [This is the big winner for us] Fact Verification: Every response comes with citations, so you can easily verify the credibility of the information and explore further if needed.

How can Perplexity help you?

  • Resource discovery. It can suggest effective keywords or phrases to refine your searches in the library catalogue or other online databases. (Stay tuned for our upcoming blog post on smart keyword searches)
  • Study Aid: It can quickly clarify topics, provide concise explanations, or further explore topics for better understanding, all of which can save time during study sessions.
  • Fact-Checking: Perplexity can quickly verify claims or statistics for essays or presentations, ensuring your work is both accurate and credible.

Some Things to Remember About Perplexity

  • Limited Conversation Depth: Perplexity isn’t as good as ChatGPT at maintaining contextual follow-ups or extended conversational exchanges.
  • Reliance on External Sources: While citations are a strong feature, the tool’s accuracy depends on the quality of its referenced sources. Always cross-check critical information where possible.
  • No Creative or Open-Ended Outputs: Perplexity isn’t designed for tasks like creative writing, brainstorming, or exploratory discussions—ChatGPT is much more suited to these tasks.

Putting it All Together.

Perplexity AI is a powerful tool for students and researchers, offering quick access to up-to-date information and helpful citations for fact-checking and further study.

Unlike ChatGPT, which excels in conversations, Perplexity delivers concise answers, making it ideal for academic work. It also helps you to discover further resources by suggesting effective search keywords.

However, it lacks the conversational depth of ChatGPT, and its accuracy depends on the quality of the sources that it cites. By combining both tools, you can make the most of your study time and make your research process more focused and effective.

AI and the Library Week three. Prompt writing: How to write effective prompts for better AI results

AI can be a valuable tool to help you to get the most out of your time in the library. By utilising the conversational style of AI tools like ChatGPT, you can collaborate with the AI to refine searches, get personalised recommendations, and quickly discover relevant resources that meet your specific academic needs.

To get the best results from AI, it’s important to ask the right questions, and this is a skill in itself,

Pixabay

This skill, known as prompt building or prompt engineering, involves structuring your queries in a way that directly impacts the usefulness of the AI’s responses. Mastering this technique can significantly improve your results from AI tools. (For more on the importance of effective prompt building, see our AI and the Library Guide here).

What is an effective prompt?

A good AI prompt is concise, structured, and specific. You might like to think of it as a formula, where each element plays a role in making the prompt clear, targeted, and focused on the desired outcome.

The elements of the formula might look something like this:

Task + Topic + Structure + Style + Level

Let’s have a look at what we mean by those elements and how we might apply them in a library scenario.

The Task component of the formula specifies what you want the tool to do. Some examples might include: Find; Summarise; Explain; Describe; Compare.

The Topic defines the subject matter or scenario that we want the Task to explore. Examples might include: The Origins of the Second World War; Romanticism in English Literature; The Aims of Positive Psychology.

The Structure specifies the format in which the response should be presented. You could ask for responses in: One Sentence; 200 words; A Bullet Point List; A table; A Graphic or Chart.

The Style specifies how the content should be written. Styles might include: Formal; Informal; Academic; Witty.

The Level of detail indicates the depth and scope of the information needed. This level of detail could be a Basic Overview or an In-Depth Analysis (or anywhere inbetween!)

Here’s an example that you might use in the library. You want to find some library resources that will help you to answer the following question: “Analyse the themes and characteristics of English Romanticism in the poetry of William Wordsworth and Samuel Taylor Coleridge.”

  • Task: Find
  • Topic: Books that discuss English Romanticism and the poetry of William Wordsworth and Samuel Taylor Coleridge
  • Structure. A bullet point list
  • Style. Academic
  • Level. An introductory overview

The full prompt may look something like this:

Please find me some academic books that give an introductory overview of English Romanticism and the poetry of William Wordsworth and Samuel Taylor Coleridge and show them in a bullet point list

This was the output that the AI provided* (for this example, we used perplexity.ai, a tool we will be reviewing in our next AI blog post):

*We always recommend checking any AI outputs for accuracy.

By getting to grips with the art of prompt building, you can more effectively communicate your needs to the AI. This ensures that the responses are both academically relevant and that they meet your specific learning needs – saving you time spent searching. You could then visit Primo, the library catalogue to see if the library has the suggested titles available for you.

As demonstrated in our example, a well-structured prompt can help you uncover valuable academic resources that can quickly help you get a better understanding of topics.

AberSkills. Your skills hub

Do you want to develop your academic writing skills, learn about using the library and its resources, get to grips with referencing, or improve your employability skills?

Good news! These topics and more are covered in the Semester 1 AberSkills programme, which is available for free for all students at Aberystwyth University.

AberSkills Workshops take place throughout the academic year and are a mixture of face to face and online sessions. Most sessions are offered in both Welsh and English.

All workshops are listed on the AberSkills website. Take a look to see what’s available and book your place in a click.

If you miss a session and want to catch-up, academic skills and library workshop teaching materials for 2023-2024 are available on Blackboard under Organisations. 2024-2025 workshop teaching materials will be uploaded soon after the session.

Protecting your research: avoiding publication scams 

Journal hijacking and fraudulent journal sites are becoming an increasing problem for journal authors, publishers and readers. Publishing scams aim to exploit researchers, promising quick publication but charging excessive publication fees. These sites can often be a clone of an established journal, set up to extract fees from unsuspecting authors. 

Publishers are becoming more increasingly aware of the problem and have taken steps to combat this new trend. Scopus academic database had 67 hijacked journals on its database by 2023 (Challenges posed by hijacked journals in Scopus – Abalkina – 2024 – Journal of the Association for Information Science and Technology – Wiley Online Library ) To help alleviate this problem, Scopus removed URLs to all journal homepages that it indexes, although the problem persists (Retraction Watch, 2023 Elsevier’s Scopus deletes journal links following revelations of hijacked indexed journals – Retraction Watch

Many authors and readers are unaware of this practice and the following resources might be useful to consider. 

Evaluating journals: 

 
Support from your library: 

 
Contact us: librarians@aber.ac.uk