AI evidence extraction

Do Early Career Researchers Consider AI as an Opportunity or a Threat? A Pathfinding Study

David Nicholas, David Clark, Abdullah Abrizah, John Akeroyd, Eti Herman, Jorge Revez et al. · Learned Publishing · 2026

AI-generated evidence extraction, verified across multiple analytical personas. Not a substitute for the peer-reviewed original.

9/10

Relevance

Evidence

Citations

0.00

FWCI

This is an AI analysis. Read the peer-reviewed original at the publisher: https://doi.org/10.1002/leap.2068

Methodology & findings

Study design

Qualitative interview study using purposive sampling.

Sample

N = 60, 1 group

Primary method

Qualitative analysis using AI-assisted thematic identification (NotebookLM) with human verification and correction by national interviewers. No quantitative statistical methods reported in abstract.

Main result

The study found that "AI is a double-edged sword which has huge potential as well as posing significant challenges." The research revealed that while "much of relevance was volunteered in answering related AI questions" regarding employment prospects, early career researchers perceive both opportunities and threats from AI. Additionally, "the AI-assisted analysis proved effective at identifying broad themes, though human oversight was essential to capture nuance, differences between cohorts, and unusual cases."

Reports effect sizes.

Research paradigm

Qualitative interpretivist

Author conclusions

The authors conclude that "AI is a double‐edged sword which has huge potential as well as posing significant challenges" for early career researchers. They also note that "the AI‐assisted analysis proved effective at identifying broad themes, though human oversight was essential to capture nuance, differences between cohorts, and unusual cases."

Risk of bias

Selection bias: purposive sampling rather than random selection may not represent all early career researchers; Sample size bias: small sample (60+ participants) acknowledged by authors as limiting generalizability; Volunteer bias: reliance on volunteered responses about job security rather than direct questioning; Country/subject representation: diversity across six countries and subjects may introduce heterogeneity in interpretation; AI tool bias: initial analysis performed by NotebookLM may introduce algorithmic biases before human review; Sample size: described as 'select and relatively small'; Geographic limitation: only six countries represented; Researcher bias in AI-assisted analysis requiring human correction; Geographic and disciplinary representation not fully specified

Limitations

The authors explicitly state that "given that we were working with a select and relatively small sample to inform a larger study, the data should be seen as illuminating and filling a research lacuna, rather than a definitive result in a fast-changing field." The study employed a "purposive and diverse sample" rather than random sampling, limiting generalizability
Additionally, the employment security topic "were asked 50 plus questions during interviews, none were directly asked about changes to job security and employment prospects," meaning data on this critical issue was entirely incidental.

Open questions raised

The authors identify that 'widespread media speculation suggests that it is entry-level positions that will be hit hardest by AI' but direct empirical evidence from early career researchers themselves on job security and employment prospects remains scarce. They position their work as filling this research gap while acknowledging the need for larger definitive studies in this rapidly evolving field.
The authors note the research fills "a research lacuna" regarding AI's impact on early career researchers' work and careers, particularly in light of media speculation about entry-level positions being hit hardest by AI.

Data: not_statedCode: not_statedExtracted from: pdf

Explore related topics

Responsible AI Use Human-AI Collaboration Mixed Methods