Natural and social science professionals n.e.c.

compiles and analyses quantitative and qualitative data, prepares reports and presents results to summarise main findings and conclusions

AI tools like Power BI Copilot, Claude, or Python-based NLP pipelines can substantially automate data compilation, thematic analysis, and report drafting, with humans mainly reviewing conclusions and ensuring accuracy.

analyses results and writes up results of tests and experiments undertaken

AI tools like Python with pandas/statsmodels, or Jupyter AI assistants, can automate data analysis and generate draft write-ups, but interpreting results in context and ensuring scientific rigour requires human oversight.

writes journal articles, books, and addresses conferences, seminars and the media to reveal research findings

AI tools like ChatGPT or Grammarly can accelerate drafting journal articles, structure arguments, and improve clarity, but original scientific insight, peer credibility, and media communication still require the human researcher.

designs tests and experiments to address research objective and find solutions

AI can suggest experimental designs and flag statistical power issues using tools like JMP or Python-based simulation libraries, but the creative and domain-specific judgment required to design valid experiments remains human-led.

applies models and techniques to medical, industrial, agricultural, military and similar applications

AI models (e.g. TensorFlow, scikit-learn) can accelerate the application of computational techniques across sectors, but translating model outputs into real-world applications still requires significant domain and contextual expertise.

presents results of scientific research to sponsors, addresses conferences and publishes articles outlining the methodology and results of research undertaken

AI tools like Gamma or ChatGPT can help draft presentations and structure conference papers, but effective scientific communication, audience engagement, and defending findings under scrutiny remain distinctly human skills.

designs and develops an appropriate research methodology in order to address the research objective

AI can suggest methodological frameworks and flag common design flaws using tools like Elicit or ResearchRabbit, but developing a rigorous, novel methodology appropriate to a specific research question requires expert judgment.

provides technical advisory and consulting services

AI tools like ChatGPT or Copilot can help draft technical reports and synthesise background information, but nuanced advisory services require contextual expertise, professional accountability, and client relationship management.

advises government, private organisations and special interest groups on policy issues

AI can rapidly synthesise evidence and draft policy briefings using tools like ChatGPT or Consensus, but advising on policy requires political awareness, stakeholder trust, ethical judgment, and accountability that AI cannot provide.

plans, directs and undertakes research into natural phenomena

AI can assist with literature reviews and hypothesis generation, but planning and directing original research into natural phenomena requires deep domain expertise, intuition, and strategic judgment that AI cannot replicate.