AI tools can generate factually incorrect information or invented sources, so-called hallucinations, which is why every output must be critically reviewed. There is a risk of scientific misconduct if AI-generated content is adopted without being labelled or passed off as an original contribution. Data protection problems can also arise if unpublished research data or personal information is entered into external systems. Another methodological problem is that AI models do not perform any real source criticism and can reproduce biases from training data.

Recommendation: Use AI tools in addition to traditional research instruments - not as a substitute!

The following tools make it easier to get started with research by helping you to find topics and develop research questions. They suggest relevant topics and questions and thus help to start the research process in a targeted manner. Well suited for creating outlines and supporting the search for terms for the literature search.

• ChatGPT

ChatGPT can support you in a scientific context during literature research by structuring topics, explaining central concepts and specifying research questions. It also helps you to formulate, revise and structure seminar papers or exposés and to summarise complex specialist texts. It can also explain methodological approaches or provide guidance on technical programming issues.

• Claude AI

Claude AI supports you above all by analysing and summarising complex scientific texts, PDFs and data sets, which speeds up literature research considerably. It helps with structuring papers, brainstorming research questions and formulating scientific texts by optimising or stylistically adapting drafts. Thanks to the large context windows, entire papers can be uploaded to get feedback on logic, argumentation or citation styles, and the tool helps you identify similarities and contradictions between multiple sources.

• Copilot

Copilot efficiently supports you in literature research by finding relevant academic studies, creating summaries and helping to structure complex topics. In Microsoft, it can create drafts for term papers or presentations in Word and PowerPoint, optimise texts and format references. The AI assistant also helps you understand difficult scientific texts by explaining concepts and suggesting structured outlines.

• Le Chat Mistral

Le Chat Mistral is a European alternative to well-known AI tools and is GDPR-compliant. In addition to its use in research and structuring scientific work, Le Chat Mistral is strong in data analysis. For example, it can execute Python code to clean data sets, perform statistical analyses and create visualisations.

• Lumo (Proton)

This privacy-compliant AI assistant (zero-access encryption) is particularly suitable in higher education, as conversations are not stored or used to train the models. This is ideal for handling sensitive research data or unpublished manuscripts. Lumo supports you by helping to analyse, summarise and structure research documents. The ability to upload documents such as PDFs and organise them into "projects" means that complex scientific texts can be processed efficiently.

• Perplexity

Perplexity.ai supports you in your academic studies with AI-supported, citation-based research that provides reliable sources and scientific literature directly. In the special "Academic" mode, you can clarify research questions, have literature clearly summarised and speed up your entry into complex topics. The tool also helps with structuring work, brainstorming and analysing academic studies by uploading documents.

• Consensus

Consensus supports you by using AI to search for answers directly in over 200 million peer-reviewed specialist publications and summarising them in a comprehensible way. You can ask complex research questions and receive evidence-based results with direct citations, which significantly speeds up literature research for seminar papers or theses. The tool also helps to identify scientific consensus on a topic, compare methodologies and export citations correctly.

• Elicit

Elicit supports you with your seminar paper or Bachelor's/Master's thesis by using AI to find relevant literature via semantic search, even without exact keywords. It automatically analyses and summarises research papers, methodologies and results, which can save you up to 80% of time on literature searches and reviews. The tool also extracts specific data from papers, creates comparison tables and helps to explore research questions.

• ORKG ask

ORKG ask supports you in the academic field by using AI to extract targeted answers to complex research questions from over 80 million scientific publications. It enables a semantic search that goes beyond mere keywords and helps you to gain a quicker overview of the state of the art in research. The tool also facilitates literature management through integrated export functions and helps to compare research contributions across disciplines.

• SciSpace

SciSpace supports you with an AI-supported platform that allows you to search, understand and summarise scientific literature more quickly. The "Copilot" allows you to highlight complex papers, formulas or tables and have them explained directly in the browser, which is particularly helpful for literature research. You can also upload your own work, use paraphrasing tools for the writing process and generate citations automatically.

• Scopus AI

Scopus AI supports you in your scientific work by answering complex research questions by analysing the Scopus database in natural language and precisely summarising relevant literature. It facilitates entry into new topics, identifies research gaps and key literature and visualises connections between different research fields. It also helps to reduce research time by providing reliable summaries based on verified content.

Access for external users

• Semantic Scholar

Semantic Scholar supports you in your academic studies with an AI-supported literature search that searches over 200 million scientific works and efficiently highlights the most relevant sources. Thanks to functions such as "TLDR" summaries (Too Long; Didn't Read) and the analysis of citation contexts, you can understand content more quickly and identify important research papers. In addition, personal libraries, Zotero integration and "Research Feeds" allow you to structure literature searches and keep them up to date.

• ChatGPT

ChatGPT is an AI writing tool that provides particularly effective support in structuring papers, formulating complex research questions and overcoming writer's block by generating outline drafts. As a writing assistant, it also helps to paraphrase academic texts, improve their style, check for errors and format the bibliography. He can also provide feedback on rough drafts, making the writing process more efficient.

• Grammarly

Grammarly supports you by checking English texts in real time for grammar, spelling, punctuation and style to improve clarity and precision. In addition, the integrated plagiarism checker and suggestions for a more formal tone of voice help you to adhere to academic standards and prepare work professionally. It is a valuable tool for refining your own writing style and overcoming language barriers.

A writing assistant that focuses on grammar correction, spell checking and tone optimisation.

• Jasper

Jasper supports you in particular by quickly creating outlines, formulating precise sections and overcoming writer's block. It helps to maintain a scientific tone, reformulate sentences and structure texts in terms of content, which makes the entire writing process more efficient. In addition, longer texts can be written consistently by using document workflows. The tool also offers functions such as plagiarism detection and formatting aids.

• Neuroflash (D)

Neuroflash supports you in particular through the AI-based structuring and formulation of texts by helping to optimise academic writing style or create outlines. Functions such as "Paraphrase" and "Rewrite" allow complex issues to be formulated more precisely and in the appropriate tone (academic/objective). The document editor also allows drafts to be developed more quickly, which makes the writing process more efficient, for example for seminar papers or theses.

AI tools can make your work much easier and support you in your scientific literature research. However, the quality of the results depends heavily on precise input, the so-called prompt. The clearer and more detailed your prompt is, the better the AI's response. Only with the right prompt can the full potential of the AI be utilised.

A good prompt is precise and clear, contextualised and goal-oriented!

It reduces room for interpretation, precisely defines the goal and framework conditions and makes quality requirements explicit.

Prompting tips

1. Define role - Specify the role in which the AI model should respond

   Example: Act as a professor of Mechanical Engineering specialising in thermodynamics

2. Provide context - Describe framework conditions such as study course, level, goal or project framework

   Example: The text is for a seminar paper in the 5th semester Bachelor Mechanical Engineering at a technical university.

3. Narrow down the topic - Formulate the topic clearly and specifically

   Example: Topic: Comparison of the efficiency of heat pump systems in industrial applications.

4. Define goal and output - Determine what exactly you need (e.g. outline, analysis, literature review)

   Example: Create a structured outline with brief explanations of the content of each section.

5. Specify format and scope - Determine the structure, length, citation style or form of presentation

   Example: Scope: approx. 800 words, factual-scientific style, with bullet points and clear section headings.

6. Formulate quality criteria and limitations - Define methodological requirements or delimitations

   Example: Consider current technological developments, name typical key figures and point out possible limitations. No speculative statements.

Example prompt:

Act as a professor of Mechanical Engineering specialising in thermodynamics.
The text is for a seminar paper in the 5th semester of a Bachelor's degree in Mechanical Engineering at a technical university.
The topic is the comparison of the efficiency of heat pump systems in industrial applications.
Create a structured outline with brief technical explanations for each section.
Length approx. 800 words, factual-scientific style with clear headings.
Take into account relevant key performance indicators (e.g. COP, seasonal performance factor), current technological developments and technical and economic limitations.