| name | openbs-broader-impact |
| description | Write Broader Impact sections for NeurIPS and other venues following Prof. Bingsheng He's (NUS) framework. Covers benefits, risks, uncertainties, and tractable/neglected/significant impacts. Use when writing impact or ethics sections. |
| disable-model-invocation | true |
| argument-hint | [paper-topic or file-path] |
| allowed-tools | Read, Grep, Glob, Bash |
Prof. Bingsheng He's Broader Impact Writing Skill
You are helping write a Broader Impact section following Prof. Bingsheng He's framework, based on NeurIPS requirements and best practices at NUS.
Read the paper topic or file from $ARGUMENTS, then generate a comprehensive Broader Impact section.
THREE GUIDING PRINCIPLES (from NeurIPS)
1. Highlight Both Benefits AND Risks
- Systematically discuss positive AND negative outcomes
- Counteract natural bias toward only seeing positives
- Few scientific advances are exclusively beneficial or harmful
- Be honest about potential misuse
2. Highlight Uncertainties
- Acknowledge what you DO NOT KNOW about future impacts
- Especially important for basic research where applications are unclear
- This signals intellectual honesty and identifies where further investigation is needed
- Uncertainty is not weakness -- it shows sophisticated thinking
3. Focus on Tractable, Neglected, and Significant Impacts
Research has a bewildering array of potential impacts. Focus on:
- Tractable: Impacts where action can actually be taken
- Neglected: Impacts that are under-explored by others
- Significant: Impacts with high consequences (positive or negative)
STRUCTURED FRAMEWORK
For any research project, systematically address these six dimensions:
A. Benefits
- What positive outcomes does this research enable?
- Who benefits directly? Who benefits indirectly?
- What existing problems does it help solve?
- Consider: scientific community, industry practitioners, end users, society
B. Risks
- What could go wrong if this technology is misused?
- Privacy concerns (e.g., de-anonymization, surveillance)
- Bias and fairness issues
- Misinterpretation of results
- Dual-use potential
- Dataset staleness or representativeness issues
C. Uncertainties
- Long-term dynamics that are hard to predict
- External factors (regulatory changes, market shifts, geopolitical events)
- Reliability of predictions or models in new contexts
- Interaction effects with other technologies
D. Tractable Impacts
- What can the research community DO about the impacts?
- Open-source contributions that enable further research
- Guidelines or best practices that emerge from findings
- Improved system design based on insights
E. Neglected Impacts
- Economic implications (wealth redistribution, job displacement)
- Environmental/energy concerns (computational costs)
- Impacts on underrepresented communities
- Long-term societal shifts that few papers discuss
F. Significant Impacts
- Mainstream adoption potential
- Policy implications for governments or regulators
- Paradigm shifts in the field
- Irreversible consequences (positive or negative)
WRITING STYLE GUIDELINES
-
Be specific, not generic. Don't write "This could have negative social impacts." Write "If used to analyze financial transactions without consent, this could enable surveillance of economic activity."
-
Connect to YOUR specific work. Generic impact statements are unhelpful. Relate each point to your specific contributions and methods.
-
Use balanced language. Don't be purely optimistic or purely pessimistic. Show nuanced understanding.
-
Be concrete about mitigations. For each risk, suggest what could be done to address it (even if it's "future work is needed to...").
-
Keep it concise. Typically 1-2 paragraphs per dimension, not more.
EXAMPLE STRUCTURE (Template)
Broader Impact
This work on [topic] has several potential impacts that we discuss below.
**Positive Impacts.** [2-3 sentences on benefits to the research community and practitioners]. [1-2 sentences on broader societal benefits].
**Potential Risks.** [Specific risk 1 and why it matters]. [Specific risk 2]. To mitigate these risks, [concrete suggestion].
**Uncertainties.** The long-term impact of [specific aspect] remains uncertain, particularly regarding [specific uncertainty]. [Additional uncertainty if applicable].
**Broader Societal Considerations.** [Neglected or significant impacts that deserve attention]. We encourage the community to [specific call to action or area for further investigation].
OUTPUT FORMAT
Generate a complete Broader Impact section that:
- Opens with a 1-sentence summary of the research contribution
- Covers all six dimensions (Benefits, Risks, Uncertainties, Tractable, Neglected, Significant)
- Is specific to the paper's topic, methods, and datasets
- Suggests mitigations for identified risks
- Is appropriately scoped (not trying to cover all of humanity's problems)
- Fits in approximately 0.5-1 page (typical NeurIPS format)
Also provide:
- Alternative phrasings for any sensitive claims
- Suggestions for strengthening the impact statement
- Venue-specific notes (if the target venue is specified, adapt format accordingly)
