Chemists

Impact

Chemists are using machine learning to predict molecular properties, propose synthetic routes and prioritise candidates before anything is made, through tools such as Schrödinger, Synthia and the predictive features in SciFinder and Reaxys. Language models draft reports, write analysis scripts, summarise literature and help with regulatory documentation, while electronic lab notebooks such as Benchling structure data for analysis. In well-funded labs, automated synthesis and high-throughput platforms are beginning to close the loop between prediction and experiment. The day-to-day shifts toward designing experiments, interpreting results and running the automation rather than doing every step by hand.

Risk

Elevated transformation: desk-based analysis and drafting automate; experimental judgement, safety and interpretation stay human.

Chemists are in the official very high exposure tier with moderate task applicability and observed usage, placing the score in the elevated band with a 2-6 year window. Literature review, data analysis, report and regulatory writing, routine spectral interpretation and early-stage candidate screening are being automated or heavily assisted now. Bench synthesis, method development, troubleshooting a failed reaction, handling hazardous materials and deciding what a result actually means remain human, although lab automation will gradually take over routine repetitive procedures in high-volume settings. The role tilts toward experimental design, oversight of automated systems and scientific interpretation, with fewer hours of manual data handling and writing per project. Quality-control and analytical roles built on repetitive testing are more exposed than research and development.

Sector readiness

Advanced in Pharma, Uneven Elsewhere

Pharmaceutical and large chemical companies have invested heavily in computational chemistry, machine-learning property prediction and automated laboratories. Smaller firms, academic labs and contract testing laboratories use generative AI for writing and analysis but have adopted the specialised tools more slowly. Adoption is rated medium-high, with a wide gap between the leaders and the rest.

1. 01

   Learn the predictive tools in your field. Retrosynthesis, property prediction and docking tools are becoming standard in discovery and process chemistry. Being fluent in them keeps you in the planning conversation.

2. 02

   Script your data handling. Python with chemistry libraries lets you process spectra and assay data quickly and understand what the AI tools are doing underneath.

3. 03

   Keep your hands skilled. Automation handles routine steps, but setting up, troubleshooting and scaling reactions still need a chemist who knows the bench.

4. 04

   Write with AI, verify everything. Draft reports and regulatory text with language models, then check every number, structure and citation. Fabricated references are a real hazard in chemistry.

5. 05

   Own the interpretation. Models propose; chemists decide. Make your value the judgement about what a result means and what to do next.

6. 06

   Move away from repetitive testing. If your role is mostly running standard analyses, seek work in method development, R&D or automation oversight before the pressure arrives.

1. 01

   Machine-learning property prediction. Models predict solubility, toxicity, activity and other properties, cutting the number of compounds that need to be synthesised and tested.

2. 02

   AI retrosynthesis. Tools such as Synthia and the predictive features in Reaxys propose synthetic routes that once took days of literature work.

3. 03

   Generative drafting and analysis. Language models write reports, regulatory sections and analysis code, and summarise literature quickly.

4. 04

   Laboratory automation. Automated synthesis platforms and high-throughput screening, combined with AI planning, are closing the loop between prediction and experiment.

5. 05

   Electronic lab notebooks and structured data. Platforms such as Benchling make experimental data machine-readable, enabling the models that use it.

6. 06

   Pharma R&D economics. Pressure to cut discovery timelines and cost drives heavy investment in computational and automated chemistry.

Pharmaceutical discovery

The most advanced setting for AI-driven design and automated synthesis; chemists work closely with computational teams and oversee automated workflows.

Industrial and specialty chemicals

Process optimisation and formulation use machine learning increasingly, but plant-scale work and safety keep chemists hands-on.

Analytical and quality-control laboratories

Repetitive testing and reporting are the most exposed; automated instruments and AI data review reduce the staff needed per sample.

Academic research

Generative AI accelerates writing and literature review; computational chemistry is well established but bench research remains labour-intensive.

1. 01

   Computational chemistry literacy. Familiarity with property prediction, docking and retrosynthesis tools is now expected in discovery roles; vendor training and online courses cover the basics.

2. 02

   Scientific programming. Python with RDKit and data libraries lets you handle data and understand model behaviour; start with your own datasets.

3. 03

   Experimental design and method development. Deciding what to test and how is where chemists add value as prediction gets cheaper.

4. 04

   Spectral and data interpretation. Judging whether a result is real, artefactual or misleading remains a human skill that automation depends on.

5. 05

   Laboratory automation operation. Knowing how to program, run and troubleshoot automated platforms is a growing requirement in industry.

6. 06

   Safety and regulatory knowledge. Handling hazardous materials and meeting regulatory standards remain the chemist's responsibility and cannot be delegated to software.

AI-guided candidate selectionExample 1

How

A discovery team uses property-prediction models to narrow thousands of virtual compounds to a few dozen for synthesis.

Gain

Fewer compounds are made and tested, shortening the discovery cycle.

Retrosynthesis planningExample 2

How

A process chemist runs a target through retrosynthesis software to generate route options, then evaluates feasibility and cost.

Gain

Route scouting that took days of literature work is done in hours.

Automated report draftingExample 3

How

An analytical lab uses a language model to draft method and results sections from structured notebook data, with chemists verifying before sign-off.

Gain

Reporting time falls and chemists spend more time on method development.

Medical Laboratory TechniciansMore exposed · exposure 43

AI impact

Standardised testing and result reporting are highly automatable through instruments and AI review.

Work moves to

Running clinical tests and quality assurance.

BioengineersDifferent skills, growing · exposure 56

AI impact

AI accelerates design in bioengineering and demand grows for people who integrate computation with wet-lab and device work.

Work moves to

Biological systems design and engineering.

Chemical EngineersComplementary, less exposed · exposure 38

AI impact

Plant design, scale-up and process safety keep chemical engineering closer to physical systems and less exposed.

Work moves to

Process design, scale-up and plant operations.

Our report is one reading of the evidence. This section is the other dataset: what people who do or know this work make of it. Nobody has scored this role yet. Sign in to add yours.