Society / Social Change
Track social change, shifting values, public sentiment and cultural transformation through structured summaries built from curated sources.
Konrad Kording | Helping Human Scientists do Better Science @ Vision Weekend Puerto Rico 2026
Summary
Human scientists frequently ask poorly defined questions, which leads to research that lacks impact. Many scientists struggle to formulate questions that are both clear and meaningful, resulting in studies that do not contribute significantly to their fields. The integration of AI into scientific workflows is often seen as a potential solution to these issues.
However, the effectiveness of AI in enhancing scientific inquiry is questionable. AI systems tend to generate ideas that replicate existing research rather than produce original insights. This raises concerns about the quality of research outputs and the potential for AI to reinforce existing biases in scientific inquiry.
The reliance on AI to streamline scientific processes may inadvertently prioritize quantity over quality. Many scientists feel pressured to publish frequently, which can lead to the acceptance of low-quality research. Without a robust framework to evaluate what constitutes 'good science,' the introduction of AI could exacerbate existing problems in the scientific community.
Efforts to improve scientific inquiry must focus on defining what good science entails. A checklist of criteria for successful research could help guide scientists in their work. Additionally, incorporating friction into the research process may encourage deeper thinking and more rigorous methodologies.
Perspectives
short
Proponents of AI in Science
- Argues that AI can help scientists formulate better questions
- Claims AI has the potential to streamline research processes
- Highlights the need for innovative tools to assist in scientific inquiry
Critics of AI in Science
- Questions the effectiveness of AI in producing quality scientific outputs
- Denies that AI can replace the need for rigorous question formulation
Neutral / Shared
- Notes that many scientists struggle with vague questioning
- Observes that the publication process often prioritizes quantity over quality
- Acknowledges the existence of tools that utilize AI in research
Metrics
other
100 ways how people get it wrong ways
common failure modes in scientific writing
Identifying these failures can help improve scientific rigor.
I have seen science fail in probably about 100 ways.
other
the first drug candidate was produced by an AI drug candidate
AI-generated drug development
This represents a significant milestone in AI's application in science.
the first drug candidate was produced by an AI.
Key entities
Timeline highlights
00:00–05:00
Human scientists often ask poorly defined questions, leading to research that lacks impact. The integration of AI into scientific workflows may not yield the expected benefits and could overlook critical thinking.
- Human scientists often ask vague questions, resulting in research that lacks impact. This underscores the importance of formulating precise inquiries in scientific work
- Even with well-defined questions, scientists may reach incorrect conclusions due to flawed methodologies. This inconsistency raises doubts about the trustworthiness of scientific results
- Integrating AI into research could address some issues, but challenges remain. AI may struggle to distinguish between significant and trivial studies because of biases in existing literature
- Many accepted scientific papers lack practical utility, while innovative concepts are frequently rejected. This indicates that the peer review process may not effectively assess research quality
- Using AI to streamline scientific workflows might not produce the expected benefits. There is a need for AI to promote critical thinking and introduce necessary friction in research
- The speaker proposes a product that encourages scientists to critically assess their ideas instead of merely automating tasks. This strategy aims to improve scientific output by fostering deeper analysis and creativity
05:00–10:00
Human scientists often formulate vague questions, resulting in research that lacks impact. The integration of AI in research raises concerns about its effectiveness, as many AI-generated ideas replicate existing human research rather than offer originality.
- Human scientists often formulate vague questions, which results in research that lacks impact and highlights a critical gap in the scientific process
- The integration of AI in research raises concerns about its effectiveness, as many AI-generated ideas tend to replicate existing human research rather than offer originality
- The pressure to publish can lead to an increase in low-quality papers, indicating a lack of mechanisms to filter out poor ideas and risking the dilution of scientific literature
- To enhance scientific outcomes, AI should introduce friction in the research process, prompting researchers to critically assess their ideas and methodologies
- A clear checklist defining good science is essential for guiding AI in evaluating research quality, as the absence of metrics complicates the distinction between good and bad science
- The speaker has created a tool called plan your science.com to assist researchers in refining their ideas, emphasizing the importance of user feedback for improving its effectiveness
10:00–15:00
Human scientists frequently pose vague questions, which undermines the impact of their research. This highlights a significant flaw in the scientific process that may hinder meaningful advancements.
- Human scientists often ask vague questions, leading to research that lacks impact and revealing a significant flaw in the scientific process