In a striking testament to artificial intelligence’s transformative grip on academia and industry, a new study reveals that 84% of researchers now incorporate AI tools into their workflows, up dramatically from 57% just one year ago. This surge, detailed in Wiley’s second annual ExplanAItions report, underscores a rapid evolution in research practices, driven by AI’s promise of enhanced efficiency amid a sobering “reality check” on its limitations.
The global survey of 2,430 researchers, conducted in August 2025, highlights AI’s tangible benefits. An overwhelming 85% report improved efficiency, while nearly three-quarters (75%) note boosts in both the quantity and quality of their output. Specific applications in research and publication tasks have jumped from 45% to 62%, with tools aiding everything from data analysis to manuscript drafting. Mainstream platforms like ChatGPT dominate, used by 80% of adopters, though specialized research assistants lag at just 25% awareness.
Yet, this enthusiasm is tempered by recalibrated expectations. Last year, researchers believed AI surpassed human performance in over half of potential use cases; now, that figure has plummeted to under one-third, averaging 30%. A key driver? Hands-on experience exposing AI’s flaws. Concerns over inaccuracies and “hallucinations”—fabricated outputs—have risen to 64% from 51%, while privacy and security worries climbed to 58% from 47%. As one anonymous researcher quipped in the study, “AI is a powerful assistant, but it’s no replacement for critical thinking.”
Barriers persist, particularly around support and training. Only 41% feel their organizations provide adequate AI resources, and 57% cite a lack of guidelines as the top obstacle to wider adoption. Corporate researchers fare better: 58% access employer-provided tools, compared to 40% overall, and they perceive AI outperforming humans in 50% of tasks—far above the global average. This disparity suggests that institutional investment could unlock AI’s full potential, reducing reliance on free, general-purpose tools favored by 70% despite 48% having paid options available.
Jay Flynn, Wiley’s EVP and General Manager for Research & Learning, captures the moment’s nuance: “We’re witnessing a profound maturation in how researchers approach AI as surging usage has caused them to recalibrate expectations dramatically. Wiley is committed to giving researchers what they need most right now: clear guidance and purpose-built tools that help them use AI with confidence and impact.” Indeed, 73% of respondents look to publishers for ethical guardrails to navigate pitfalls like bias or intellectual property risks.
The implications are profound. As AI integrates deeper into the research lifecycle, it democratizes complex tasks—62% now see it excelling in error detection, plagiarism checks, and citation organization. Yet, without addressing the “guidance gap,” adoption risks stalling. Full findings, due in late October, promise deeper insights into discipline-specific trends.
Looking ahead, optimism endures. Researchers view AI not as a panacea but a vital ally in accelerating discovery. In 2025, the question isn’t whether to use AI, but how to wield it responsibly. As adoption hits 84%, the research world stands on the cusp of an AI-augmented renaissance—one grounded in realism, not hype.
Leave a Reply