evidoria

← Back to browse

Good practice Imported

New Jersey's AI Task Force — Policy Synth for Citizen-Sourced AI-Workforce Policy

United States of America · Trenton · See the United States of America profile

Evidence: Descriptive / self-reported Top 25% 73/100 · Ask Evidence Copilot about this practice

New Jersey's state AI Task Force used the open-source Policy Synth AI-agent system, combining automated research with input from 2,200+ workers and a 5,000-strong public-servant survey, to generate 2,552 proposals narrowed to 4 recommendations.

1,451
AI-generated policy proposals (8-week process, by April 2025)
20
Issue areas covered by AI-generated proposals
2,200+
Private-sector workers engaged via All Our Ideas (3 weeks)
1,101
Worker-sourced proposals
15
Issue areas covered by worker-sourced proposals
5,000
Public servants surveyed
73 %
Public servants wanting to learn more about AI
2,552
Combined AI- and worker-sourced proposals
40
Proposals shortlisted
4
Final recommendations delivered to state leadership
3,000+
State job descriptions reviewed for degree requirements
250,000 USD
GitLab Foundation grant (January 2025)
500,000 USD
GitLab Foundation grant (2026, for expansion)
New Jersey's AI Task Force — Policy Synth for Citizen-Sourced AI-Workforce Policy

Details

Maturity
Pilot
Promoter
New Jersey Artificial Intelligence Task Force (Office of the Chief AI Strategist) with Northeastern University's Burnes Center for Social Change and Citizens Foundation
Period
2023-2026
Keywords
AI policy, labour market, public-sector workforce, government innovation

Context

Following Governor Phil Murphy's October 2023 executive order establishing an AI task force, New Jersey's Office of the Chief AI Strategist worked with Northeastern University's Burnes Center for Social Change and the Icelandic non-profit Citizens Foundation to apply Policy Synth, an open-source framework orchestrating teams of AI agents around public problems, to how AI is changing the state's workforce.

Objectives

Generate and prioritise policy proposals on AI's effect on New Jersey's workforce by combining AI-agent-driven research with citizen and public-servant engagement.

Activities

Over an eight-week process completed by April 2025, Policy Synth's agents ran automated research across academic literature and grey sources to generate 1,451 policy proposals spanning 20 issue areas. A separate 'All Our Ideas' public engagement exercise collected input from over 2,200 private-sector workers over three weeks, and a survey reached 5,000 public servants (73% said they wanted to learn more about AI), yielding a further 1,101 worker-sourced proposals across 15 issue areas. Human subject-matter experts reviewed the AI-generated material at each stage, narrowing the combined 2,552 proposals to 40 shortlisted options and then to 4 recommendations delivered to state leadership.

Results

Two concrete actions followed: free, self-paced generative-AI training for all New Jersey public servants (built with InnovateUS) and a commissioned AI-powered labour-market monitoring tool. A related application, launched with a further $250,000 GitLab Foundation grant from January 2025, used Policy Synth to review over 3,000 New Jersey state job descriptions for unnecessary degree requirements; a second $500,000 GitLab Foundation grant awarded 2026 is funding expansion to three to five additional states.

Conclusions

No independently measured before/after outcome (e.g. actual change in hiring of workers without degrees) had yet been published for either strand.

Implementation
Indicative cost
Medium (€50k–€500k)
Time to results
Medium (1–3 years)
Staffing & skills
Office of the Chief AI Strategist (Beth Simone Noveck), Northeastern University's Burnes Center for Social Change, Citizens Foundation (Policy Synth), Human subject-matter experts reviewing AI-generated proposals at each stage

Conditions for success

  • Open-source Policy Synth framework for AI-agent-orchestrated research
  • Parallel citizen/worker engagement (All Our Ideas platform, statewide survey)
  • Human expert review gating AI output at each narrowing stage
  • External philanthropic funding (GitLab Foundation) for follow-on applications

Common failure modes

  • No independently measured before/after outcome published for either the training program or the job-description review strand

Commonly funded by

National / regional programmes Philanthropic / foundation funding

Indicative funding routes for practices of this type — always check each programme's current calls and eligibility rules.

Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.

Data sources

Where this practice's information was retrieved from, and when.

Attachments

Similar practices you may find useful