AI pioneer Geoffrey Hinton, Ph.D., who won the Nobel Prize in 2024, recently quipped: “I think for mundane intellectual labor, AI is just going to replace everybody.” While Hinton’s take may not represent a consensus view, generative artificial intelligence is already changing many jobs and industries. The World Economic Forum’s Future of Jobs Report 2025 estimates that 92 million jobs will be displaced by AI by 2030. The report also estimates that AI will create 170 million jobs in that time, a net growth of 78 million jobs, or 7% of today’s total employment. Additionally, Goldman Sachs reported that AI could automate 300 million jobs globally.
AI-related changes are already evident in the R&D space and the broader STEM sphere. For example, AI Engineer and AI consultant were at the top of LinkedIn’s 2025 Jobs on the Rise report. And some senior AI leaders earn millions in salary with Meta reportedly offering $100 million pay packages to a few elite candidates, according to WSJ..
As in many other industries, AI is disrupting basic R&D tasks, automating low-seniority jobs while promising to create new roles. Here are five R&D jobs whose tasks can be largely automated, and five new careers that might emerge.
5 R&D jobs at risk of AI-based automation
1. Research assistants
Roles involving routine data analysis and reporting are likely to be displaced by AI. Research assistant roles often involve data collection, cleaning and entry, which can all be increasingly automated by agentic AI features such as Deep Research tools from frontier AI companies as well as bespoke tools from academic publishers (though accuracy is still uneven). In addition to searching academic databases, AI can summarize papers and create annotated bibliographies. Generative AI can suggest experimental designs, track variables, and automatically populate electronic lab notebooks with data, eliminating the need for RAs to manually collect and record data or draft experimental procedures.
2. Junior data analyst
Junior data analysts are often tasked with creating basic reports and visualizations, which can now be done by tools such as Microsoft Fabric Copilot, ThoughtSpot and ChatGPT data analyst plugins, or even in integrated development environment (IDE) software or in Jupyter notebooks, including Google Colab. GenAI tools can automatically connect to data sources, clean data and run statistical analysis. In addition, Natural language generation tools can write summaries of insights and conclusions from datasets.
3. Process specialists
Process specialists who run A/B tests to optimize prototypes are also likely to be displaced by AI automation. Generative AI can automatically create multiple variants from past patterns or content rules and run real-time experiments with live adjustments. Algorithms can shift traffic towards better-performing variants while learning from others and use Bayesian optimization to find the best combination of variables. When the test is complete, AI can automatically employ the best solution.
4. Simulation engineers
GenAI-based systems can act as surrogate models for finite element analysis or computational fluid dynamics, where neural networks can predict simulation results in milliseconds. For instance, tools such as NVIDIA’s physicsnemo, DeepONet, and BOSS (Bayesian Optimization Structure Search) are already available. Such AI can also automate or assist in meshing. In addition, more traditional machine learning models can predict optimal mesh density, refine the mesh where needed, and even generate synthetic mesh topologies, while genAI systems can also automate the analysis of the large data sets produced by FEA and CFD. Generative design software, like Autodesk’s Dreamcatcher, can integrate with FEA to automate prototype optimization and other design tasks based on simulation results.
5. Clinical study coordinators
Low-seniority roles in clinical trial coordination focused on lower-level data entry and processing are at risk of being replaced by AI. These roles involve document management, protocol monitoring and reporting logistics, which can be managed by eClinical platforms that are powered by AI. NLP-based AI tools can screen potential participants through surveys via emails and phone calls. AI can automatically track and update documents such as consent forms that require participant signatures. AI assistants can schedule appointments, send reminders and optimize staff allocation. AI tools can collect and transfer data and provide reports and dashboards.
5 R&D jobs AI can create (or already has)
1. AI research scientist
While the origins of artificial intelligence date back to the late 1950s, it wasn’t until decades later that research institutions ranging from IBM to SRI International and Xerox PARC launched formalized efforts to develop practical machine intelligence. Now, scientists focused on AI are conducting research with decidedly more practical leanings. These scientists, distinct from data scientists, are more likely to design the architecture of frontier AI models. That is, while a data analyst largely asks, ‘what happened,’ and a data scientist is more likely to probe into ‘why did it happen,’ an AI research scientist is more focused on inventing new ways for machines to perform. AI research scientists develop the foundation models and tools that domain experts can explore in their respective fields. For example, a biologist might use AI models like AlphaFold3 or Boltz-2 to predict protein folding and then design a wet-lab experiment to verify the prediction.
2. AI research data curator
Before the wave of interest in genAI, poor data quality cost organizations across sectors an average of $12.9 million annually, according to a 2020 Gartner stat. But as AI automates data collection and transfer, while sometimes hallucinating, human curators will be needed to oversee this process. Data curators will be experts who will ensure the quality and relevance of data pipelines that feed AI models. They will be responsible for maintaining clean, labeled datasets for use in ML models. Any data that cannot be automatically collected by AI will have to be inputted by scientists in this role. Already, universities like Stanford, for instance, have posted positions with titles like research data analyst 2 and research data manager with competitive six-figure salaries.
3. Computational discovery dngineer
This role will be responsible for training neural networks to make accurate predictions through simulations. This will help the process of automating simulations with AI to be faster and more accurate. These experts will have skills in programming, surrogate modeling and relevant simulations. As AI automates FEA and CFD simulations, simulation engineers may find themselves in roles more similar to this, training and overseeing AI. Already, professionals working in this broader field are bridging two worlds, spanning traditional simulation tools (ANSYS, ABAQUA, COMSOL), ML frameworks like TensorFlow and PyTorch, languages like Python and C++ and specialized knowledge in physics-based neural networks.
4. AI auditor
This role will be responsible for validating AI-generated findings to ensure they are ethical and reproducible. Their key skills will include statistical testing, regulatory knowledge and AI expertise. This will be an important role to ensure AI results are unbiased and safe. Factors driving interest in AI auditing include the European Union’s AI Act, which has fines of up to €40 million or 7% of worldwide turnover, and interest in AI ethics at frontier AI labs and Big Tech companies.
5. Synthetic data designer
Gartner’s 2021 prediction that 60% of AI training data would be synthetic by 2024 (up from 1% in 2021) has largely proven accurate. This role will use AI to simulate data when real data is unable to be acquired. They will use professional judgment to verify AI outputs from generative models. They will need to know how to train AI models to generate datasets. This can be useful in modeling rare events, as data is scarce in this field. According to Emergen Research, companies such as NVIDIA, IBM, Microsoft, Google and Amazon are already using synthetic data. Emergen reported that the global synthetic data generation market size was $1.3 billion in 2023.
