AI in Academia: General Information

Prompt - A query put into an AI model to obtain a specific output or answer. The query can be a sentence, question, detailed instructions for a task, or a set of keywords.

Large Language Model (LLM) - An AI tool resembling a chatbot that produces natural language texts, trained on vast amounts of data across various subjects and fields. These models can perform diverse tasks, engage in dialogue with end users to understand queries and improve responses, adjust linguistic register as needed, and more. Some even provide sources for their responses.

Machine Learning - A subfield of AI technology focusing on training artificial intelligence systems to learn continuously, improving gradually without human intervention or explicit programming.

Deep Learning - A type of machine learning that aims to mimic neural networks of the human brain to train AI tools to perform complex tasks, relying on pattern recognition and visual or textual predictions.

Librarians  

Please contact an expert librarian from the library staff list

For general inquiries: 

Tel: 08-6461111/2

Email: yaatz@bgu.ac.il

Whatsapp    

Facebook  

Mail: P.O.Box 653, Beer-sheva 8410501, Israel

Academic Databases: Current language models do not have access to paid academic databases. They have access to free online databases, some of which are academic. It's necessary to critically assess secondary sources obtained from AI tools.

Privacy: All information input into a language model is used for its future learning. It's important to note that the information we share here is "exposed" to some extent.

Hallucinations: A known problem with many language models whereby if they lacks reliable sources, they invent information that appears to be entirely possible, but is in fact false. It's essential to verify any and all responses received against reliable sources.

Biases: Language models are known for social, gender-based, cultural, ethnic, and other biases. All content suggested by these tools should be evaluated critically and according to specific needs.

Data Quality: Language models are trained on vast amounts of data the sources of which are not available to the general public. Therefore, do not assume that the quality of content received from the model is necessarily good. It is recommended to double-check all AI-produced content before use.

Educated Search: Use AI tools for broad-scope searches that require extracting and merging information from many diverse sources to save you time and web browsing.

Prompt: Ensure a precise prompt for your query. Define in advance the search goal, content register, target audience, and response length to help the language model focus and provide the optimal answer.

Dialogue: Language models continue to learn constantly, so it's advised to provide feedback on the received response and refine the query if necessary. It's possible to actually "converse" with the tool to get better results, so do that.

Transparency: Keep in mind that not all fields, tasks, courses, or instructors will be open to the use of artificial intelligence tools. You should verify with the instructor what use is permitted (if any) in advance.