Is being a Biostatistician
at risk from AI?

Not in the foreseeable future, but the role will shift. AI is already automating routine descriptive statistics, standard tests, and table generation—tasks that once consumed 30-40% of a biostatistician's time. However, regulatory agencies require human accountability for statistical decisions, and the judgment needed to design trials, choose appropriate methods for complex endpoints, and defend analyses to the FDA cannot be delegated to AI. The biostatisticians most at risk are those doing purely execution work (running pre-specified analyses on clean data). Those leading study design, regulatory strategy, and interpretation of ambiguous results remain essential.

Routine automation is happening now—AI coding assistants already generate R and SAS code for standard analyses faster than manual coding. Over the next 2-3 years, expect AI to handle most descriptive statistics, common modeling tasks, and first-draft SAP sections. The 3-5 year horizon brings better handling of missing data strategies and sensitivity analyses, but complex adaptive designs, causal inference in observational studies, and regulatory negotiation will remain human-led for the next decade. The shift is toward higher-level work, not elimination of the role.

It depends on your career direction. If you work in early-stage drug discovery, genomics, or precision medicine, ML skills (especially for high-dimensional data and predictive modeling) are increasingly valuable. If you focus on late-stage clinical trials and regulatory submissions, deep expertise in causal inference, Bayesian methods, and adaptive designs offers more resilience—these are areas where interpretability and regulatory acceptance matter more than predictive accuracy. The safest bet is to stay fluent in both traditional biostatistics and modern computational methods, but prioritize the skills your specific therapeutic area and company stage demand.

Salaries for senior biostatisticians with regulatory experience and strategic design skills are likely to remain strong or even increase, as AI makes their judgment more valuable by handling grunt work. Entry-level and mid-level roles focused on executing pre-defined analyses may see wage pressure, as AI reduces the labor hours needed for those tasks. The market is bifurcating: high-value strategic roles (designing trials, leading regulatory interactions) versus commoditized execution roles. Geographic arbitrage may also intensify, with companies using AI to reduce reliance on expensive US-based analysts for routine work.

Yes, but not impossible. Entry-level roles that once involved learning by doing routine analyses are shrinking, as AI handles those tasks faster. New biostatisticians need to demonstrate higher-order skills earlier—understanding study design rationale, regulatory context, and cross-functional communication. Internships and roles that expose you to protocol development, SAP authoring, and regulatory submissions are more valuable than pure programming or analysis positions. Consider targeting smaller biotech firms or CROs where you'll wear multiple hats and gain strategic exposure faster than in large pharma's siloed analyst roles.

Academic biostatisticians face different pressures. Grant-funded research roles often involve novel methodological development, which AI currently cannot do, offering more insulation. However, service roles (consulting on colleagues' studies, running standard analyses) are highly automatable and may see reduced demand or lower billing rates. The academic job market's emphasis on publications and methods innovation provides some protection, but teaching roles may also shift as AI tutors handle more introductory statistics instruction. Industry roles tied to regulatory submissions have clearer human-accountability moats than academic service work.

The ASA's PStat certification signals professional credibility but doesn't directly address AI resilience. More valuable are credentials that demonstrate regulatory expertise (e.g., RAC from RAPS if you want to pivot toward regulatory affairs) or specialized methods training (Bayesian adaptive designs, causal inference). Practical experience leading FDA/EMA submissions, serving on Data Safety Monitoring Boards, or publishing methods papers in your therapeutic area builds a reputation AI cannot replicate. Focus on credentials that open doors to strategic, high-accountability work rather than technical skill badges that AI will soon match.