4.3. Which AI tools and models are relevant?

There are many AI tools. Some AI tools can be useful to use during your studies. Below is an overview of commonly used AI tools, including their applications, advantages, and limitations. Note: this is just a small selection of the many AI tools available.

AI-tool

Application Advantage Limitation

1. Copilot

Text, coding, image Broadly applicable, intuitive Hallucinations, no source citation

2. Claude

Text, brainstorming, reasoning, image, coding Strong in nuance, long context Limited availability

3. ChatGPT

Text, brainstorming, reasoning, image, coding Broadly applicable, intuitive interface Prone to errors, hallucinates sources

4. Elicit

Literature research Systematic and scientific, summaries with source citation English-language focus

5. Consensus

Evidence-based answers, scientific consensus Peer-reviewed sources Less suitable for creative tasks

6. NotebookLM

Literature analysis, working with own documents Contextual search in own sources Limited to uploaded data

7. DALL-E, Adobe Firefly

Image generation Creative visual output Limited control over style

It is important to know that the above AI tools are based on underlying LLMs such as GPT-4 (OpenAI), Claude 3 (Anthropic), Gemini (Google). Be aware that these models do not understand the world like humans do. They do not know what is factually valid, but are statistical and based on probability.

Watch the video below for a demonstration of how to search for sources in the following AI tools: Consensus, ScholarGPT, Elicit, and Semantic Scholar: