When Cambridge researchers pulled together more than 3,000 scientific papers on climate change mitigation, they uncovered a stubborn gap between rigorous science and actionable policy—and it starts with how scientists frame their own recommendations. Dr. Vangelis Danopoulos and his team at the Statistical Laboratory found that while the underlying research was solid, the policy recommendations aimed at turning findings into real-world action were often treated as an afterthought, buried at the end of papers or divorced entirely from what the data actually showed.
This matters because the world is running out of time. Scientists face mounting pressure to deliver solutions for net zero, yet trust in institutions sits at historic lows. If the bridge between laboratory findings and boardroom decisions crumbles under the weight of vague language or unrealistic demands, neither policymakers nor the public will know what to believe—or what to do.
The team created a tool called Evidence Communication Rules for Policy (ECR-P) to assess the quality of policy recommendations in papers published since 2019 on green energy and transportation. After screening thousands of papers, they conducted deeper analysis on 23 and identified three recurring problems. First, levels of uncertainty were rarely disclosed or even highlighted, leaving policymakers blind to the risks embedded in recommendations. Second, advocacy language—words like "must be forbidden"—crept into supposedly objective research, blurring the line between science and campaigning. Third, and perhaps most revealing, many recommendations read like wish lists, floating free from the study's actual findings and rooted instead in whatever happened to be topical at the time.
The consequences are real. Policymakers often lack deep expertise in any single field; they navigate broad, complicated terrain across energy, transport, infrastructure and more. When they receive recommendations muddied by uncertainty, emotions or wishful thinking, they're building policy on unstable ground. "Being clear about the uncertainties, being clear about the trade-offs is so important if we're going to bring our science to the people, industries and governments who will be charged with turning that science into decisions," Danopoulos explained. "Highlighting what we don't know is just as important as highlighting what we do know."
The researchers did find bright spots. One paper examining how to roll out electric vehicles across the United Kingdom treated policy recommendations as a research outcome from the start, not an optional afterthought—and the resulting guidance was notably sharper and more grounded. That model suggests a path forward, though Danopoulos is clear that no single tool can fix the problem alone. Scientists need training in how policy actually gets made. Funding bodies should require rigorous policy reporting as a condition of grants. Institutions must reshape expectations around what constitutes quality research output.
The work sits within the broader AI for Net Zero project, which aims to make research more accessible and useful to the complex machinery of policymaking. The message isn't to scold scientists but to acknowledge reality: the world doesn't need more disconnected research or more earnest wishes. It needs clarity, honesty about what we don't know, and recommendations that land where they can actually be used.