Science and Research Content

Elsevier report highlights data quality gaps in R&D decision-making -

Elsevier’s Corporate R&D of the Future report finds that 63% of respondents say their organizations have made important research and development (R&D) decisions using data or evidence later found to be inaccurate, incomplete, or outdated. The findings are based on a global survey of more than 1,000 R&D leaders, managers, and researchers working at R&D-led corporations.

Although 76% of respondents expressed confidence in the quality of their data, 51% said their organizations had been required to make important R&D decisions before all the necessary evidence was available. The findings highlight gaps between confidence in available information and the reliability and completeness of evidence used in decision-making.

The survey also examined the adoption of artificial intelligence (AI) in R&D. While 81% of respondents expect AI to fundamentally transform R&D decision-making over the next five years, only 47% said AI is currently used regularly within their organization's R&D function. Another 40% reported that AI applications are being piloted.

Stuart Whayman, president of Corporate Markets at Elsevier, said that confidence in data does not necessarily indicate its accuracy. He noted that R&D teams face pressure to make decisions quickly, often without complete evidence, and that confidence cannot compensate for information that is incomplete, outdated, or unreliable. He also observed that AI could increase the consequences of data-quality issues by amplifying the evidence on which systems rely. According to Whayman, organizations that can consolidate their evidence and trace and verify the information underlying their decisions will be better positioned to benefit from AI.

The report examines how R&D teams use data and evidence in decision-making, the sources they rely on, the extent to which information is connected, and how AI is being adopted and evaluated.

According to the findings, 78% of respondents have access to the third-party evidence they need. However, 56% of those with access still reported at least one barrier to using it. Across all respondents, the most frequently identified challenges were scattered information sources (42%), delays in finding and synthesizing evidence (32%), and a lack of awareness of available evidence (30%). Access to third-party evidence was highest among respondents in Asia-Pacific (APAC), at 90%, compared with 77% in Europe and 73% in North America.

AI adoption is increasing, although regular organizational use remains limited. While 47% of respondents said AI is used regularly in their organization's R&D function, another 40% reported that it is being piloted. At the individual level, respondents most commonly use AI for data analysis (71%), literature searches (70%), and competitive intelligence (70%).

The perceived importance of reliable information in AI applications was also reflected in the survey. Overall, 86% of respondents believe organizations that combine AI with trusted, high-quality evidence will outperform those that do not.

Compliance, usability, and accuracy were among the leading factors influencing respondents' willingness to use AI. Some 55% said they would be more willing to use AI systems that comply with applicable laws and regulations, while 54% identified ease of use and integration with existing workflows as important considerations. An equal proportion cited factual accuracy and safety. Regulatory compliance was particularly important among European respondents, 64% of whom said it would increase their willingness to use AI.

The findings also indicate differing views on the challenges of implementing AI at scale. Nearly half of respondents (48%) said large-scale AI implementation would present a significant challenge. Over the next five years, the most frequently identified organizational priorities were developing AI skills across the workforce (65%), converting data and evidence into actionable insights (62%), and adapting AI to organizational knowledge and workflows (59%). Respondents in North America were more likely to identify large-scale implementation as a significant challenge (56%), compared with those in Europe (50%) and APAC (34%).

For literature searches, general-purpose AI applications were more commonly used than specialized scientific AI tools. Among respondents who answered this question, 48% reported using general-purpose tools such as ChatGPT or Copilot, while 23% used specialized scientific AI tools. For data analysis and modeling, specialized tools were the most commonly selected option.

The survey also identified regional differences in how R&D leaders and managers assess evidence quality and AI adoption. In North America, 42% said evidence that was not sufficiently robust or reliable reduced their confidence, compared with 32% in Europe and 23% in APAC.

Views on the challenges of implementing AI at scale also varied by region. The proportion of R&D leaders and managers who considered large-scale AI implementation a significant challenge was 52% in North America, 43% in Europe, and 28% in APAC.

AI-generated summaries and outputs were considered essential evidence by 61% of respondents in APAC, compared with 36% in Europe and 30% in North America. Regulatory compliance was another area of regional variation. Among R&D leaders and managers, 62% of European respondents said compliance with laws and regulations would increase their willingness to use AI, compared with 47% in APAC and 43% in North America.

The regional analysis included 183 R&D leaders and managers from North America, 234 from Europe, and 275 from APAC. Respondents who selected "Don't know" were excluded, resulting in slight variations in the sample sizes across individual questions. South America (34 respondents) and the Middle East and Africa (33 respondents) were excluded from the regional comparison because of their smaller sample sizes.

The full Corporate R&D of the Future report includes additional comparisons by region, organization size, and professional role, along with recommendations for R&D organizations. It is available on Elsevier's website: https://www.elsevier.com/insights/corporate-r-and-d-of-the-future.

Click here to read the original press release.

STORY TOOLS

  • |
  • |

sponsor links

For banner ads click here