Article reader Listen + reading controls
Article reader
Preparing the reader…
Reading settings
ReDDDoT and the Missing Infrastructure of Responsible Innovation¶
Funding responsible-technology research is necessary. Funding isolated projects is not sufficient.
The more consequential opportunity in the National Science Foundation’s Responsible Design, Development, and Deployment of Technologies (ReDDDoT) program is to build the infrastructure of a field: people, methods, shared evidence, institutional arrangements, and translation pathways that make responsible practice repeatable beyond a single grant.
This distinction matters because the market systematically underproduces that infrastructure. Companies can capture value from products; they struggle to capture the full value of public test methods, community capacity, interoperable documentation, negative results, and long-term study of societal effects. Government and philanthropy can invest where commercial incentives are weak—but only if they treat the portfolio as more than a collection of worthy projects.
The NSF ReDDDoT program was created with philanthropic partners to support research, workshops, education, and translation across fields including AI, biotechnology, quantum information, robotics, advanced communications, cybersecurity, energy, and disaster mitigation. The program explicitly seeks broad communities of interest, workforce development, community participation, and pathways to benefit while avoiding harm.
Those goals imply a field-building strategy.
The Field Has an Incentive Problem¶
Responsible-technology work frequently produces outputs that are socially valuable and organizationally homeless:
- a method for evaluating disparate impacts;
- a community-generated set of design requirements;
- a taxonomy of harms;
- a dataset documenting failure conditions;
- a participatory design protocol;
- an assurance-case template;
- or a longitudinal study showing that a deployment changed behavior in an unexpected way.
Academic incentives favor novel publication. Product teams favor features and delivery. Compliance functions favor demonstrable adherence to current rules. Community organizations are often funded for participation in a project rather than for sustained technical capacity. No actor naturally owns the work of maintaining, standardizing, and translating the result.
The grant ends; the method decays.
A public-interest program should deliberately finance the connective tissue that keeps knowledge usable.
From Projects to Shared Assets¶
Every ReDDDoT award can produce project-specific findings. The portfolio should also produce shared assets that later teams can adopt and improve:
- reference methods for stakeholder and community participation;
- common vocabularies connecting social harms to technical and organizational mechanisms;
- evaluation protocols for accessibility, equity, privacy, safety, and accountability;
- machine-readable documentation and traceability patterns;
- synthetic or permissioned benchmark datasets;
- open training curricula and case studies;
- model contract language for responsible procurement;
- incident and near-miss taxonomies;
- and repositories of negative results and design decisions.
These assets need stewardship. A repository of grant outputs is not infrastructure if no organization validates, versions, documents, and supports their use.
NSF can ask projects to contribute to shared standards and fund maintainers or centers that integrate results across awards. Philanthropic partners can support the participation of civil-society and community organizations whose value is not well captured by conventional research funding. Professional societies can turn mature methods into durable practice.
Community Participation Requires Community Capacity¶
Inviting marginalized or economically disadvantaged communities into technology design does not equalize power. Universities, agencies, and companies arrive with lawyers, engineers, institutional funding, and control over the agenda. Community organizations may be asked to supply trust, lived experience, recruitment, and legitimacy on a short timeline.
Responsible field-building should invest in the community side of the partnership:
- direct and sufficient compensation;
- independent technical and legal expertise;
- control over how community-generated data and knowledge are used;
- governance roles rather than advisory status alone;
- long-term relationships beyond a proposal period;
- and the ability to publish disagreement or withdraw.
Capacity changes participation from extraction to co-production. It also improves the research. Communities that understand the technology and retain their own evidence can challenge assumptions throughout deployment rather than contributing only during discovery.
Translation Needs Boundary Organizations¶
Research does not move automatically into practice. Responsible-technology methods often require adaptation to a particular sector, procurement process, data environment, or regulatory regime.
Boundary organizations can bridge those contexts. They may be university centers, public-interest technology organizations, standards bodies, federally funded research and development centers, professional societies, or sector-specific consortia. Their function is to translate among research, community knowledge, engineering, law, policy, and operations.
They can:
- validate methods across multiple settings;
- help product teams operationalize values as requirements and tests;
- train practitioners;
- maintain open tools;
- convene affected communities and implementers;
- and aggregate lessons that individual organizations cannot disclose publicly.
ReDDDoT’s workshop and translation-oriented award structures are useful precisely because field formation requires relationships and institutions, not only research findings.
Philanthropic Partnership Should Add More Than Money¶
The collaboration among NSF and philanthropic partners creates an opportunity for complementary roles. Federal science agencies provide scale, public accountability, research infrastructure, and rigorous merit review. Philanthropies can often move quickly, tolerate organizational experiments, fund advocacy and community capacity, and support work that falls outside traditional academic categories.
The partnership should preserve clear governance. Funders must disclose how priorities are set, protect the independence of merit review, manage conflicts of interest, and avoid allowing a narrow set of private institutions to define the public interest.
Done well, blended funding can support the full ecosystem:
- NSF funds research and translational evidence;
- philanthropy funds participation, capacity, and institutional experimentation;
- government agencies provide use cases and pathways to public deployment;
- industry contributes technical access and operational environments;
- and communities hold authority over values, impacts, and acceptable tradeoffs.
No single actor should dominate the evidence.
Responsible Technology Needs a Workforce, Not a Specialist Function¶
The field will not scale if responsibility remains the job of a small ethics team reviewing everyone else’s work. Engineers, product managers, designers, data stewards, acquisition professionals, lawyers, executives, and operators all make decisions that shape risk and public value.
Training should therefore be role-specific and practice-based:
- engineers learn to translate harm scenarios into architecture, telemetry, and tests;
- product leaders learn to frame outcomes, value tensions, and stop conditions;
- acquisition teams learn to demand evidence, data rights, and update controls;
- executives learn to assign decision rights and accept residual risk explicitly;
- operators learn system limits, intervention, incident reporting, and appeal processes;
- community partners gain the technical fluency and resources to exercise real influence.
Education is successful when it changes the work product—not when participants can repeat a list of principles.
Measure the Portfolio as an Ecosystem¶
The 2024 awards eventually invested more than $18 million across 44 teams.1 Outputs and publications will matter, but the portfolio should also be evaluated through network effects:
- How many methods are reused outside the originating project?
- Which tools or curricula remain maintained after grant funding?
- Did community organizations gain durable capacity and governance power?
- Do awardees collaborate across technology areas and sectors?
- Which results influence procurement, standards, professional practice, or deployed systems?
- How quickly can a new team adopt prior evidence rather than begin from zero?
- Are negative findings preserved and discoverable?
- Did the portfolio change who participates in technology development?
These measures reflect whether ReDDDoT is creating an ecosystem or merely funding episodes.
The Strategic Inference¶
Responsible innovation is a collective-action problem. Each organization benefits when trustworthy methods, trained people, representative evidence, and informed communities exist. Few organizations can justify building all of that infrastructure alone, and some benefit privately from avoiding scrutiny.
Public and philanthropic funding can correct the imbalance. Its highest leverage is not in producing one more framework. It is in creating shared capacity that changes the economics of responsible practice—making credible evaluation, participation, traceability, and learning easier to adopt and harder to omit.
The lasting success of ReDDDoT will not be the number of technologies labeled responsible. It will be whether future teams inherit a richer field of evidence, institutions, and people capable of making responsibility part of ordinary design and deployment.
This essay was substantially revised in July 2026 to replace the original call-for-proposals summary with an evidence-based field-building analysis. It incorporates award information published after the original post.
Organizations building the operating mechanisms behind responsible AI may find related material in my portfolio and research. For a practical discussion, send a direct inquiry.
References¶
- U.S. National Science Foundation, “NSF 24-524: Responsible Design, Development, and Deployment of Technologies (ReDDDoT).”
- “NSF Inviting Proposals for 2024 Edition of Responsible Design, Development & Deployment of Technologies Program,” ExecutiveGov, January 2024. This report was the historical prompt for the original post.
-
U.S. National Science Foundation, “NSF and philanthropic partners invest more than $18M to prioritize ethical and societal considerations in the creation of emerging technologies,” September 23, 2024. ↩