Today, the pharma sector, like many others, appears to be on the cusp of a transformation thanks to the influence of automation on drug discovery careers. Companies ranging from Insilico Medicine to Big Pharma firms like Janssen and Sanofi are putting AI at the heart of their operations. Many are increasingly recruiting workers from the tech sector.
Complex dynamics are at play amidst this shift. McKinsey acknowledged that more than 60 trillion U.S. work hours in the pharma sector could be impacted from technologies such as machine learning. But the integration of automation in the pharmaceutical industry is not simply a story of job loss and displacement. While 2023 will go down as notable year for layoffs in the sector, AI also promises to redefine existing jobs as create new ones as well.
Barry Burgdorf, partner in corporate and finance at Hogan Lovells, has an optimistic outlook on this shift. “The popular press often focuses on questions like, ‘Will ChatGPT take away jobs?’ or ‘Will scientists become unemployed because of AI in research and scientific medical research?'” Burgdorf believes the answer to that is ‘no’, at least for a considerable time. But AI does have significant potential across the drug discovery and development lifecycle. “When we examine every stage of drug development — starting with a medical scientist’s initial idea, progressing to IP protection, the formulation of a potential drug candidate or molecule, conducting toxicity and efficacy studies, through preclinical work, and even extending to the marketing and distribution of that drug, there’s not a place that that AI can’t play a role,” added Burgdorf.
An exploration of the influence of automation on drug discovery careers
To shed light on the varying susceptibility to automation of different job roles within the pharmaceutical industry, we created the stacked bar chart below that examines thirteen roles in the industry, each evaluated on fourteen different metrics that contribute to their risk of automation.
The visualization below was not based on a simple zero-shot output from large language models. To produce it, we manually reviewed job descriptions to identify baseline data on core skills, tools and responsibilities while also considering research data on factors that increase the risk of automation. We used large language models from OpenAI and Anthropic to assist with validating and refining the outputs, using an iterative approach to verify the output. Finally, we use Python libraries like pandas to parse the data and create the following visualization with seaborn and matplotlib.
What does automation mean for different roles in drug discovery careers?
The factors range from the level of information processing and documentation required to the potential for AI integration into the role. Each role is scored on these factors, and the total score determines its ultimate exposure of automation. Higher scores don’t necessarily equate to a higher risk of job loss, but higher scores do indicate an elevated possibility that the role could evolve or be displaced.
- Drug Manufacturing Workers: These professionals are responsible for the production of pharmaceutical products and equipment. High risk of automation as a result of the repetitive nature of tasks, high level of information processing, and documentation involved. The experience and expertise of such workers can be hard for robots to completely emulate. McKinsey singles out workers focused on labelling and assembly operations as having a high risk of automation.
- Lab Technicians: These workers perform laboratory tests and analyses. Medium risk of automation as a result of the repetitive tasks and high levels of information processing. These professionas’ work also involves complex decision-making and data analysis which are less prone to automation. The job growth for these workers is around 7%, according to BLS.
- Quality Assurance/Control Analysts: These professionals perform scientific analyses to evaluate the quality and safety of various materials and products. Medium risk as a result of repetitive tasks, high levels of information processing and the need for regulatory compliance. Such professionals require high levels of education, training and expertise relative to lower level quality-control inspectors, which are factors that are less susceptible to automation.
- Data Scientists: These professionals collect, analyze and interpret large amounts of data. These high-salaried professionals have a relatively low risk of automation. This role requires high levels of complex decision-making, creativity, strategy, data analysis, education and training. While the data analysis component of data scientists’ work is more susceptible to automation, this trend could result in job evolution rather than displacement. As automated tools can handle routine tasks, data scientists can focus on more complex, strategic challenges. According to BLA, their job outlook is much better than average.
- Research Scientists: These scientists devise, formulate and execute investigative protocols that tackle deficits in scientific knowledge. Relatively low risk of automation. This role centers around high levels of complex decision-making, creativity, strategy, data analysis, collaboration, education and training. Career growth for medical scientists is a healthy 17%, according to BLS.
- Clinical Trial Coordinators: These workers manage clinical research and assist in clinical trials. Low risk of automation as a result of high levels of complex decision-making, collaboration, regulatory compliance, education and training. But the information processing and documentation part of clinical trial coordinators’ work could be automated. BLS projects a growth rate of about 7% for clinical laboratory technologists and technicians from 2021 to 2031, which is in line with most other occupations.
- Regulatory Affairs Specialists: These professionals help a company or organization comply with the regulations that apply to their products or services. Low risk of automation as a result of to high levels of complex decision-making, creativity, strategy, collaboration, regulatory compliance, education and training.
- Clinical Research Associate (CRA): These professionals run clinical trials to test drugs for their effectiveness, risks and benefits. Low risk of automation as a result of to high levels of complex decision-making, collaboration, regulatory compliance, education and training. The information processing and documentation part of their work, however, will likely be increasingly automated. According to BLS, the related position of health information technologists and medical registrars could see 17% growth over the next decade,
- Biostatistician: These workers analyze data from medical research studies. Low risk of automation as biostatisticians’ work requires substantial levels of complex decision-making, creativity, strategy and expertise. In general, BLS projects employment of mathematicians and statisticians to grow 31% from 2021 to 2031.
- Pharmacovigilance Specialist: These professionals monitor and report the effectiveness of pharmaceutical products. Low risk of automation given the high levels of complex decision-making, training and experience required.
- Medical Science Liaison (MSL): These scientific professionals work for life science organizations such as pharmaceutical companies. Low risk given the high degree of creativity, strategy, training, expertise and human interaction involved.
- Regulatory Affairs Manager: These workers oversee the regulation process for products requiring governmental approval. Low risk given the high levels of complex decision-making and creativity required.
- Medical Information Specialist: These professionals provide information and support to patients and healthcare providers. Low risk of automation given the nature of complex decision-making and experience required. The information processing and documentation portion of their work could be automated.
Here’s a closer look at the factors and the weights used in the bar graph. ‘L’ stands for low, ‘M’ for medium and ‘H’ for high”:
|Job Role||Ult. Risk of Auto.||Info Proc. & Doc. (10%)||Repet. Tasks (5%)||Routine Data Anal. (2.5%)||Adv. Data Anal. (2.5%)||Compl. Decision-Making (10%)||Creat. & Strat. (5%)||Educ. & Train. (10%)||Collab. & Comm. (10%)||Reg. Compliance (5%)||Exp. & Expert. (15%)||Job Evol. Pot. (10%)||Resil. to Ext. Shocks (15%)||Econ. Impact (20%)||Pot. for AI Integ. (20%)|
|Drug Manufacturing Workers||L||L||L||H||H||H||H||M||M||L||M||H||M||L||H|
|Quality Assurance/Control Analysts||M||L||L||H||L||M||M||L||L||L||L||M||M||L||M|
|Clinical Trial Coordinators||H||L||M||H||M||M||L||L||L||L||L||L||L||L||M|
|Regulatory Affairs Specialists||H||L||M||H||M||M||L||L||L||L||L||L||L||L||M|
|Clinical Research Associate (CRA)||H||L||M||H||M||M||L||L||L||L||L||L||M||M||M|
|Medical Science Liaison (MSL)||H||M||H||H||L||M||L||L||M||L||L||L||M||H||M|
|Regulatory Affairs Manager||H||L||M||H||L||M||L||L||L||L||L||L||L||M||M|
|Medical Information Specialist||H||L||M||H||M||L||L||L||M||L||L||L||L||M||M|
Automation influence on drug discovery careers: The human-AI synergy
Historically, technology has tended to fuel job evolution rather than job extinction. In that vein, McKinsey foresees a growing need for technological literacy and human skills such as creativity, collaboration and communication. Professionals are not necessarily immune to automation. Earlier this year, Goldman Sachs in their report “The Potentially Large Effects of Artificial Intelligence on Economic Growth” highlighted professions like law, engineering, data analysis and administrative work as most susceptible to automation from generative AI.
In pharma, data-heavy jobs like data science are prone to transformation from automation, but also have potential for AI augmentation. But ultimately, the future of AI-based automation is not set in stone. AI replacing humans across industries is not inevitable, nor is it necessarily beneficial. It thus becomes crucial for workers in industries with high automation potential, including pharma, to take an active role in shaping how AI impacts their work. Employees should participate in redefining their roles and determining how AI tools can best augment human skills rather than replace them.
Thoughtful, strategic implementation of automation and AI can help biopharma mold the workplace of the future rather than be shaped by it. The industry may have lagged in technology adoption thus far, but the opportunity is ripe to tap AI in empowering ways that benefit both workers and companies.
A note on the methodology
In creating our automation risk assessment methodology, we considered 14 factors, each contributing to a composite score reflecting an occupation’s automation risk. Some of these elements (like repetitive tasks) have a direct relationship to the risk score, while others (like creativity and decision-making complexity) are inversely related. For instance, having a job with a high degree of repetition invites a higher degree of risk than one with considerable variety. Similarly, a job with high decision-making complexity is less prone to automation.
Some criteria such as regulatory are more complex to evaluate. While some aspects such monitoring and reporting can be automated as a result of their repetitive and data-intensive nature, other aspects such as interpretation and application of complex regulations tend to require human expertise. Given the latter, we concluded that positions with a regulatory focus tend to be more resistant to automation.
Here is the ranking criteria with a note which factors were inverted:
- Information Processing and Documentation (10%)
- Repetitive Tasks (5%)
- Routine Data Analysis (2.5%)
- Advanced Data Analysis (2.5%): Inverted. Higher scores mean lower risk. Jobs that require advanced data analysis skills, such as those of data scientists who define data workflows, are less likely to be automated. Thus, higher scores in this category signal lower automation risk.
- Complex Decision-Making (10%): Inverted.
- Creativity and Strategy (5%): Inverted.
- Education and Training (10%): Inverted.
- Collaboration and Communication (10%): Inverted.
- Regulatory Compliance (5%)
- Experience and Expertise (15%): Inverted.
- Job Evolution Potential (10%): Inverted.
- Resilience to External Shocks (15%): Inverted.
- Economic Impact (20%):
- Potential for AI Integration (20%): No inversion needed.
Filed Under: Data science, Industry 4.0, machine learning and AI