What is known and documented.

What would substantiate this

• Disclosed underlying models, vendors, or APIs (e.g., OpenAI, Anthropic, Google)
• Technical architecture documentation or whitepapers
• Engineering job postings, GitHub presence, or named technical staff
• Independent third-party technical review
• Peer-reviewed publications

What is publicly available

No technical documentation, no named model providers or infrastructure vendors, no public code repositories, no engineering staff identified on the website’s leadership materials, no peer-reviewed publications, no whitepapers.

“World’s most sophisticated” is a superlative claim ordinarily made by laboratories with published benchmarks. It does not appear to have been made by any third party in reference to Marici.

Detailed AI-claim analysis →

What would substantiate this

• An inventory of the 50 tools with descriptions of function
• Demonstrations or screen captures of any of them in operation
• Disclosure of whether tools are proprietary builds or wrappers around third-party APIs
• Evidence of internal evaluation, accuracy benchmarks, or audit

What is publicly available

Three named example tools (Intelligence Analyst, Prosecutor, Behavioral Scientist), each described in a single sentence on the homepage. No inventory of the remaining ~47 tools. No demonstrations. No technical documentation. No accuracy or false-positive rate disclosures despite the homepage claim of “predictive mapping at 80%+ accuracy.”

Detailed AI-claim analysis →

What would substantiate this

• Names of the six cities
• Names of partner law-enforcement agencies
• Baseline trafficking measurements and methodology
• Independent evaluator or academic review
• Government, NGO, or media confirmation from any of the named cities
• Court records, prosecution data, or arrest figures

What is publicly available

The cities are not named anywhere on Marici’s public-facing material reviewed. No partner law-enforcement agency is identified by name. No methodology for measuring “collapse” is published. No independent confirmation from any government, academic, journalistic, or NGO source could be located that corroborates either the geographic scope or the “second largest system on Earth” historical claim.

Independent reports from the US State Department, UN agencies, and India’s National Crime Records Bureau do not mention Marici or such a collapse.

The problem

The lifetime figure of 86,000 “lives saved” implies, on average, more than 11,000 rescues per year over the organization’s operational lifetime. The 2025 figure of 39,000+ “girls saved” is more than nine times the entire annual budget of the organization in dollars.

The most recent IRS filing reports total annual expenses of $5.48 million. That implies a cost per “life saved” of roughly $140 — yet the donor page asks $500 to “Save 1 Life.”

What “saved” appears to mean

The numbers are only internally consistent if “saved” includes prevention reach — community awareness contacts, school outreach, public campaign impressions — rather than identified individual victims removed from trafficking situations.

If that is the definition, donors are not being told. The word “rescued” is used interchangeably with “saved” across donor-facing materials, and neither term is defined.

Why this matters

This is not aspirational fundraising language. It is a specific dollar-to-outcome ratio presented at the point of donation. It implies that an individual donor’s $500 contribution will result in one identifiable life saved.

Under California charitable solicitation law and IRS Publication 1771 guidance, specific impact representations to donors should be substantiable.

What we asked

• How is “Save 1 Life” defined?
• Is $500 based on audited cost-per-outcome data?
• Who verifies these outcomes?
• Are donors informed the figure may be a projection rather than a guaranteed outcome?

No response.

What would substantiate this

• Named Stanford department, school, or institute and its public-facing page
• Principal investigator’s name and Stanford affiliation
• IRB approval number or study registration (e.g., ClinicalTrials.gov)
• Public Stanford communication confirming the partnership
• Scope of work, methodology, and publication timeline
• Whether Stanford has independently validated Marici’s aftercare or impact claims

What is publicly available

Stanford does not appear to operate a department literally named “Precision Medicine Department.” Its closest comparable units are Stanford Precision Health and the Stanford Center for Biomedical Informatics Research. No Stanford-published material referring to a Marici partnership could be located in public university communications, faculty pages, or study registries.

The partnership may be real and simply under-publicised. The claim is reproduced verbatim from Marici’s own Aftercare page; the question is one of verification, not denial.

Current status

As of the last available check (11 June 2026, via the Wayback Machine), the relaunched donate.html displayed two different EINs on the same page: EIN 85-3455940 in the US trust strip and EIN 82-1536804 in the footer tax-deductibility notice. The mismatched number, 85-3455940, cannot be matched to any nonprofit in the IRS exempt organizations database. The live site is currently behind a Cloudflare challenge and has not been re-verified since 11 June 2026.

June 2026 relaunch: both EINs on one page

A Wayback Machine capture of marici.org/donate.html, dated 11 June 2026, records a relaunched, static version of the donation page carrying three EIN-bearing elements:

• The US trust strip: EIN 85-3455940
• The footer tax-deductibility notice: “MARICI is a registered 501(c)(3) nonprofit organization, and donations are tax-deductible in the US. EIN: 82-1536804”
• The page’s own script, which sets the trust-strip value for a US visitor to EIN 85-3455940 and separately records a Canadian charity number, 82732 5150 RR0001

This is the same discrepancy first observed in May 2026, on a page that had, in the interim (as of 1 June 2026), shown only the matching EIN.

Documentation request

For full transparency we have asked Marici to confirm whether two separate EINs were ever in circulation on donor-facing materials and, if so, what donations made under the unmatched EIN have been classified as for IRS tax-deductibility purposes. No response received.

What’s missing publicly

• Publicly posted audited financial statements
• Disclosure of an independent audit committee
• Form 990 filings posted directly on marici.org (available only through third-party databases)
• Annual report
• Public list of board members beyond officers

Why it matters

For an organization with $4M+ in annual revenue making large operational and technological claims, the absence of board-level audit oversight transparency is unusual. Donors cannot independently assess whether the organization’s financial controls match the scale of its claimed activities.

What the archive shows (all captured 11 June 2026)

• about.html: “five massive interconnected cities” collapsed in 2025; “32 cities across 21 countries” for the wider deployment
• solution.html: “five massive interconnected cities” collapsed in 2025; “PROVEN 5 CITIES · 2025 COHORT”; “QUEUED 42 CITIES · 23 COUNTRIES”
• faq.html: “six major interconnected cities”; a five-year goal of “41 hotspots”

The adjective changes (“massive” vs. “major”), the count changes (five vs. six), and the wider footprint is given three different totals — 32, 41, and 42 — on pages captured within hours of each other on the same day.

What would substantiate this

• A single published list naming the cities involved in any of these counts
• A consistent, dated figure distinguishing cities completed from cities queued
• The operations data or methodology behind the five, six, 32, 41, and 42 figures
• A public explanation reconciling the same-day discrepancy between the three pages

The two totals

Marici’s FAQ page states a lifetime total of over 105,000 children freed and over 1,700 criminal networks neutralized. Claim 04 on this site documents a different set of totals, drawn from marici.org together with Marici’s GuideStar profile and Millie Giving page: over 655 networks neutralized and 39,000+ girls saved in 2025, with over 86,000 children saved since launch.

Neither page carries a publication date or version marker. Neither claim references or supersedes the other.

What’s missing

• A published methodology defining “freed” and “neutralized”
• A dated version history showing which total is current
• Reconciliation between the FAQ’s lifetime figures (105,000 children; 1,700 networks) and Claim 04’s lifetime figure of 86,000 children and its 2025 figure of 655 networks
• Independent verification of either total

What the archive shows

research.html (archived 11 June 2026) states a “0% Re-trafficking rate across the 2018–2024 completed aftercare cohorts”; aftercare.html (archived the same day) displays “RE-TRAFFICKING RATE 0%” alongside “HOMES 9 · ACROSS 5 CITIES” and “CAPACITY 340 SURVIVORS.” Neither page publishes a cohort size, a follow-up method, or an evaluation source.

What would substantiate this

• The size of each completed cohort
• The follow-up method used to determine re-trafficking status, and its duration
• A definition of what counts as a “completed” cohort
• Independent evaluation by a named academic or clinical partner

Two facts, stated side by side

Marici’s IRS Form 990 record, filed under its founding name Take Her Back, reports 0 employees for tax year 2017 and total expenses of $10,163 that year; the record also reports 0 employees for tax year 2020. The re-trafficking figure is claimed for cohorts completed from 2018. No conclusion is drawn here from placing these two facts together.

What Marici’s site says

index.html: “Marici’s 350+ interdisciplinary operatives — intelligence collectors, analysts, prosecutors, community mobilizers, AI engineers, criminologists, rescue personnel, aftercare, and government relations specialists — deploy a single coordinated system across every city we enter.”

faq.html: “We deploy 350+ interdisciplinary operatives — intelligence collectors, AI engineers, criminologists, prosecutors, community mobilizers, and survivor care specialists — in a single coordinated system across every city we enter.”

What the IRS filing says

Form 990 Part I, line 5 reports 2 total employees for tax year 2024, 2 for 2023, 4 for 2022, and 1 for 2021. The 2024 filing separately reports $248,332 in salaries and benefits and a $1,042,324 “Consultants” expense line. See the ProPublica filing pages, e.g. FY2024, FY2023, FY2022.

What’s missing

The IRS employee count covers only W-2 employees of Marici’s US entity — it does not by itself rule out a larger contracted or partner-staffed workforce. But the filings do not name who the 350+ operatives are, what country they work in, or whether they are paid, contracted, or volunteer.