2

SciPoll 1013: Which AI Models Do Experts Use and Trust?

Which of the following have you used in your scientific or professional work in the past three months?

Share Page
Post on LinkedIn
Copy URL to Clipboard
Copied!

Results
(229 Answers)

Answer Explanations

  • ChatGPT (OpenAI) Claude (Anthropic) Gemini (Google) Grok (xAI)
    user-154906
    I have used tons of different LLM models, however, the ones I used the most for scientific topics are Claude and Grok.
  • Other
    user-11803
    Openevidence
  • Claude (Anthropic) DeepSeek
    user-597773
    They are well designed based on algorithm.
  • ChatGPT (OpenAI) Other
    user-568782
    Grammar check (AJE, Editage)
    Grammarly 
  • Gemini (Google)
    user-887652
    I use AI only to the extent imposed by Google searches, and never as a final authority.
  • ChatGPT (OpenAI) Gemini (Google)
    user-441029
    I use AI only to collect primary sources for literature overview on the topic of interest, after that I do all the work myself, as this takes less efforts than checking after AI hallucinations. For that any AI will work, Gemini and Chat GPT are easiest to access, so I usually use them, I tried others - no difference.
  • ChatGPT (OpenAI) Gemini (Google)
    user-498547
    I use both ChatGPT and Gemini because they complement each other. If both give me the same answer, I think I can believe in it. Both are very useful to find information about any subject of your interest
  • ChatGPT (OpenAI) DeepSeek
    user-157767
    For editing work and getting more articles.
  • ChatGPT (OpenAI) Gemini (Google) Other
    user-96711
    AlterAI is used to consider the reliability of literature claims.
  • Copilot (Microsoft) Gemini (Google) Other
    user-460380
    I used to search on Google and Bing, and models like Gemini and Copilot (Microsoft) make my research task simpler.

  • ChatGPT (OpenAI) Gemini (Google)
    user-267027
     AI has helped me overcome language barriers and manage my time more efficiently. I first use it to build an initial understanding of a topic and gather basic information. This provides a solid foundation on which I apply my own expertise, critical thinking, and scientific judgement to develop higher-quality scientific work. 
  • None — I use AI assistants only outside of work
    user-109201
     I have not used AI in my scientific or professional work during the past three months. I currently use AI only for personal or non-professional purposes. 
  • ChatGPT (OpenAI)
    user-649290
    For language refinement it is good. 
  • ChatGPT (OpenAI) Claude (Anthropic) Copilot (Microsoft) Gemini (Google)
    user-301040
    I consider them better
  • ChatGPT (OpenAI) DeepSeek Gemini (Google) Other
    user-582664
    I have used Gemini and benefited greatly from it in solving technical problems that I have encountered. As for my use of DeepSeek, I have benefited greatly from it in medical matters. I have also used Perplexity and SciSpace in the field of scientific research.
  • ChatGPT (OpenAI) Claude (Anthropic)
    user-508016
    I utilized AI agents to enhance the drafting of my ideas, resolve coding issues, and improve graphics and tables. I did not use AI for bibliographic references, which were adequately cited within the manuscript using reference managers like Mendeley, Zotero, or BibLaTeX.

  • ChatGPT (OpenAI) DeepSeek
    user-890708
    使用了一些人工智能软件，功能更强大，日常的写作和查询资料很方便。
  • None — I use AI assistants only outside of work
    user-378118
    no me or any scientific journal trust the veracity of AI, yet - for another century perhaps if ever. 
  • ChatGPT (OpenAI)
    user-866433
    I found AI assistant useful for academic writings, verifications of formulations and calculations, implementation of idea via simulation scripts.
  • ChatGPT (OpenAI) Gemini (Google) An open-weight model I run or access directly (e.g., Llama, Mistral, Qwen)
    user-953899
    I didn’t use that much AI in my research works so far, but I tried giving AI the data to analyze the literature to review so far. I’m very  impressed AI can manage. Exhaust amount of patient data and literature data in a very short amount of time.
    Reliability is still a valid thing that many researchers are skeptical about. But with good accuracy.
  • Claude (Anthropic) Other
    user-876636
    Claude and Perplexity
  • ChatGPT (OpenAI) Gemini (Google) Other
    user-188852
    And I find them very useful 
  • Gemini (Google)
    user-139009
    Refining scientific tone: Transitioning drafts into a more formal, academic style.
    Drafting and structuring: Organizing abstract concepts or complex findings into coherent, scannable outlines.
    Grammar and flow: Editing multilingual text to meet the high publishing standards of international journals
  • Gemini (Google)
    user-951744
    Our institution has signed up for the Gemini model, which commits not to use our data to train it.
  • None — I don't use AI assistants at all
    user-813332
    I don't use AI assistants yet as I am not sure I can trust them.  I need to test them out on things I know a lot about.
  • None — I use AI assistants only outside of work
    user-487971
    I don't trust the platform as it generates several erroneous output which needs expert review which will take time to review as well as it will affect the creativity of the expert in the future
  • ChatGPT (OpenAI) Copilot (Microsoft)
    user-814722
    I have used ChatGPT and Microsoft Copilot to support scientific and professional work. These tools assisted me in my academic work
  • Claude (Anthropic)
    user-194414
    I like claude because of its ethics, but also because i has a lot f features i feel are perfect for science. claude science, cowork, projects, design, code... And you can learn a lot. 
  • ChatGPT (OpenAI) Gemini (Google)
    user-316248
    Eu usei eles para correção ortográfica e coesão apenas.
  • ChatGPT (OpenAI) Claude (Anthropic) An open-weight model I run or access directly (e.g., Llama, Mistral, Qwen) Other
    user-871577
    I have a subscription to Perplexity which I use occasionally. 
  • Claude (Anthropic) Other
    user-228600
    Internal safe institutional llms 
  • ChatGPT (OpenAI) Claude (Anthropic)
    user-516400
    I use ChatGPT to help me in writing good english, spelling, grammar and clarity in many documents and emails for scientific work. Claude has been more devoted by my questions to CVs and related documentation in job applications.
  • Copilot (Microsoft) Other
    user-377267
    I used AI developed by an agency for systematic literature review. There, search strategies are set, and all AI readouts are checked by two scientists. 
  • Gemini (Google)
    user-762565
    This is to understand certain scientific terms.
  • ChatGPT (OpenAI) Claude (Anthropic) Copilot (Microsoft) Gemini (Google)
    user-234128
    Claude seems to be the best or more reliable one. I use AI primarily to help with statistical code (R). 
  • ChatGPT (OpenAI) Gemini (Google)
    user-365645
    These are available easily 
  • ChatGPT (OpenAI) Claude (Anthropic) Copilot (Microsoft)
    user-395064
    I mainly use Chat GPT as I am used to it and can tell when the answers are suboptimal. 
  • Grok (xAI)
    user-222308
    I used it just for grammar and clarity 
  • ChatGPT (OpenAI)
    user-432588
    I used it because for scientific writing review 
  • ChatGPT (OpenAI)
    user-490306
    I have not used chatgpt during my research work and thesis write up. Because all the research work needs to be evidenced by literature via a scientifically accepted journals. Besides, thesis goes through the process of plagiarism which easily detect the AI use in the write up. But chatgpt can help in literature assistance and good academic communication. 
  • ChatGPT (OpenAI) Copilot (Microsoft)
    user-468267
    Only for brief consultations due to uncertainties.
  • None — I don't use AI assistants at all
    user-439854
    I don't trust AI assistants 
  • ChatGPT (OpenAI)
    user-746645
    I mainly use ChatGPT to help with literature searches, summarize research articles, improve English writing, and check whether manuscripts follow journal requirement. I always verify the information before using it in research or clinical work.
  • ChatGPT (OpenAI) Claude (Anthropic)
    user-851286
     As far as the design of coding and analysis in ecology is concerned, Claude Code is currently the most suitable option.

  • ChatGPT (OpenAI) Copilot (Microsoft) Other
    user-151126
    Scite
  • ChatGPT (OpenAI) Copilot (Microsoft) DeepSeek
    user-164084
    I have used some of the AI models as indicated above. They are quite handy and useful for data gathering and interactive discussions. 
  • Copilot (Microsoft)
    user-756416
    For work-related tasks, I primarily use Copilot connected to my institutional account because of the data protection and security features associated with this institutional subscription.
  • Claude (Anthropic) Gemini (Google) Other
    user-923069
    To check my English.
    I also ude SciSpace.
  • ChatGPT (OpenAI) Claude (Anthropic) Copilot (Microsoft) Gemini (Google)
    user-531244
    Models are approaching commoditization. The trend is to stable different models in gateways and apply tokens to the appropriate model for the desired work.