Responsible AI · Human-in-the-Loop

Responsible AI & Data Handling Statement

Practical, Human-Centered AI Implementation

Newbloom AI LLC

newbloomai.com

Newbloom AI helps organizations adopt AI in practical, secure, and measurable ways. Our rule is simple: AI never makes decisions without people. AI prepares the work; people make the call. Newbloom AI does not promote AI as a replacement for organizational accountability.

Foundational Philosophy

People Decide, AI Prepares

In our work, AI does the research and preparation. It gathers the facts, compares the options, and turns them into recommendations framed as simple choices: yes or no, or A, B, or C. People review every one of those choices and make the decision. The goal is to give a team many well-prepared decisions it can review quickly, not to take decisions away from it. This is especially important in government, nonprofit, public-sector, workforce development, veteran serving, and mission-oriented environments where workflows may affect people, programs, funding, compliance, or public trust.

Newbloom AI’s Implementation Loop

1Identify a repetitive, document-heavy, or administrative workflow.
2Determine where AI can safely and usefully assist.
3Have AI prepare each decision as a clear choice, and keep staff reviewing and approving every one.
4Document the workflow and review checkpoints.
5Train users on the tool, the process, and the limitations.
6Measure whether the workflow improves.

Principles

Responsible Use Principles

AI should support staff, not hide accountability.
AI generated output should be reviewed before it is relied on.
AI does not make eligibility, legal, medical, financial, personnel, or benefits decisions. It prepares them for the people accountable for them.
AI tools should be matched to the sensitivity of the workflow and the data involved.
Client policies, procurement requirements, privacy obligations, and data handling rules should guide implementation.
Staff should understand both the usefulness and limitations of AI tools.
AI workflows should remain usable, reviewable, and accessible for the staff and communities they serve.
AI workflows should be documented so they can be reviewed, improved, and governed.
Practical outcomes matter more than novelty.

High-Risk Workflow Caution

Newbloom AI does not automate high-stakes decisions. Workflows involving eligibility, benefits, healthcare, legal rights, finances, personnel decisions, public safety, or other sensitive outcomes should be handled carefully and reviewed by qualified human decision-makers. AI may still be useful for administrative support, summarization, drafting, triage, document organization, or staff assistance in those contexts — but final decisions remain with the appropriate people, governed by the client’s policies and legal obligations. The client remains responsible for determining which laws, policies, data classifications, approval processes, and security requirements apply to its workflows.

Newbloom AI LLC · newbloomai.com · Responsible AI Statement · Page 1 of 2

Responsible AI & Data Handling

newbloomai.com

Confidentiality & Access

Data Handling Practices

Newbloom AI operates within client-approved environments and follows the organization’s existing data security protocols, procurement requirements, privacy obligations, and confidentiality rules. Practical safeguards include:

Defining what data is needed for the project.
Limiting access to information based on project needs.
Avoiding unnecessary use of sensitive or personally identifiable information.
Using client-approved tools and environments where required.
Following agreed confidentiality terms.
Using written scopes of work to define project boundaries.
Applying access controls and secure collaboration practices.
Documenting workflows, tools, assumptions, and review requirements.
Reviewing outputs before delivery.
Maintaining founder-led oversight of project quality and security.

When outside specialists or subcontractors are used, access is limited to information needed for their assigned work. Newbloom AI remains responsible for project management, quality review, security expectations, and final deliverable approval.

Output Review

AI Quality Control

Reviewing AI generated content.
Checking source documents or underlying data before relying on important outputs.
Understanding when AI output is uncertain.
Avoiding overreliance on automated recommendations.
Designing workflows where staff can correct and improve outputs.
Adding human review checkpoints.
Documenting known limitations.
Measuring whether the workflow improves after implementation.

Client Fit

Policy Alignment

Reviewing existing AI or technology-use policies.
Identifying approved tools and restricted tools.
Designing workflows around client-approved platforms.
Creating staff guidance for approved use cases.
Documenting review and approval steps.
Helping teams translate policy into day-to-day workflows.

Adoption Support

Staff Training & Documentation

Responsible AI adoption requires more than access to tools. Staff need training, examples, guidance, and documentation that connect AI use to their actual work. Newbloom AI provides or develops project-appropriate materials such as:

Practical AI literacy training.
Workflow-specific training.
Prompt examples and process guides.
Human-review checklists.
Responsible-use reminders.
Manager guidance for oversight.
Documentation for repeatable AI-supported workflows.
Adoption roadmaps and next-step recommendations.

Implementation Philosophy

Start with real workflows. Use approved tools. Use AI where it clearly helps. Let AI prepare the decision; let people make it. Protect client information. Train staff. Document the process. Measure the outcome.

What This Statement Is Not

This statement is not a legal opinion, cybersecurity certification, privacy audit, compliance determination, or substitute for the client’s own policies, legal counsel, procurement office, or security review. It describes Newbloom AI’s implementation posture and the practical safeguards it uses when helping organizations adopt AI.