Compassion Benchmark

Viewing archive: Apr 18

Back to latest
Daily evidence briefing · 2026-04-18

Daily Evidence Briefing

Evidence-linked score assessments, sector intelligence, and emerging risks from overnight research across all published benchmark indexes. Each finding is sourced from primary evidence — litigation records, regulatory filings, investigative reporting, and international legal instruments.

1,155Entities scanned
15Entities assessed
6Score changes
13Scores confirmed
1 band change proposed tonight

How this works

Every night, research agents scan all 1,155 benchmarked entities for new evidence across litigation, regulatory filings, investigative reporting, and international legal instruments. Flagged entities receive full 40-subdimension assessments.

Score changes are proposals, not automatic updates. A human analyst reviews all proposals before published scores change. Confirmations — where research affirms the published score is accurate — are documented alongside changes.

Score movements

Entities with significant evidence-based score movement from overnight research. Each card is a dossier entry.

Iran

CountriesApplied0.85 confidence
2.82.5
-5.3 pts
critical

7,015 documented deaths in 2025–26 Iran massacres; EU sanctions renewed; UN crimes-against-humanity framing.

Evidence record
  1. 1
    HRANA documented 7,015 deaths including ≥6,508 protesters (Feb 5, 2026)
  2. 2
    Amnesty: greatest massacre in Iran 'in many decades'
  3. 3
    UN Human Rights Council resolution January 2026 extended fact-finding mandate
  4. 4
    EU sanctions renewed April 13, 2026; 16 new entities added March 2026
  5. 5
    Execution of six PMOI/MEK-affiliated political prisoners April 11, 2026

Ford Motor

Fortune 500Applied0.75 confidence
48.442.5
-5.9 pts
functional

1.4M F-150 recall with 18-month delay between NHTSA flag and formal filing; largest recall in scan window.

Evidence record
  1. 1
    Ford recall of 1,392,935 F-150 trucks (2015–17) for unintended downshifts, filed April 14, 2026
  2. 2
    18-month delay between NHTSA October 2024 flag and April 14, 2026 formal recall
  3. 3
    Two injuries and one accident confirmed pre-recall
  4. 4
    Largest single safety recall in Fortune 500 scan window
  5. 5
    Recall root cause: degraded electrical connections in 6R80 transmission lead frame

OpenAI

Ai LabsApplied0.5 confidence
40.638.8
-1.8 pts
functionaldeveloping

Illinois SB 3444 committee vote April 24 is pivotal governance signal; flagged for re-assessment.

Evidence record
  1. 1
    New Yorker investigation (100+ sources, secret Sutskever memos) alleges CEO Sam Altman exhibits 'consistent pattern of lying' about safety commitments; superalignment team received 1-2% compute vs promised 20%
  2. 2
    Florida AG investigation into ChatGPT's alleged role in FSU mass shooting, examining platform safety, CSAM enablement, and foreign adversary data risks
  3. 3
    Whistleblowers filed SEC complaint alleging OpenAI illegally barred employees from raising safety risks with regulators; whistleblower Suchir Balaji died by suicide
  4. 4
    DEI commitment page scrubbed from website (Jan 2025); only 16% of technical roles held by women
  5. 5
    70+ copyright lawsuits allege training on content without consent; ongoing litigation in Southern District of New York

Meta Platforms

Fortune 500Applied0.8 confidence
12.210.9
-6.3 pts
critical

$375M NM + $6M CA verdicts; Section 230 shield eroding; legislative momentum (STOP CSAM, KOSA, COPPA 2.0).

Evidence record
  1. 1
    $375M New Mexico verdict (March 24, 2026) — child safety
  2. 2
    $6M California social-media addiction verdict (March 25, 2026)
  3. 3
    STOP CSAM Act reintroduced April 16, 2026 citing Meta verdicts
  4. 4
    TechPolicy.Press analysis April 3: Platform Design Litigation Yields Historic Verdicts
  5. 5
    Senators push kids online safety after verdicts

New York City

Us CitiesApplied0.6 confidence
48.456.3
+7.9 pts
functional

REP April 6 is first citywide racial equity framework in NYC history; implementation pending.

Evidence record
  1. 1
    NYC Preliminary Citywide Racial Equity Plan released April 6, 2026 — first governmentwide racial equity framework in NYC history
  2. 2
    Mayor Mamdani announcement of REP; data-driven agency goals across 7 domains
  3. 3
    NYC published True Cost of Living Measure — standard-setting
  4. 4
    Seven-domain framework including Housing, Community Safety/Rights/Accountability, Economy, Health
  5. 5
    NYC highest-ranked US city in index makes this benchmark-setting signal

Target

Fortune 500Applied0.75 confidence
92.582.5
-10 pts
exemplary

January 2025 DEI rollback material I1/EQU/ACC event; band remains Exemplary.

Evidence record
  1. 1
    Target DEI rollback announced January 2025
  2. 2
    NAACP opposition to Target's 2025 DEI rollback
  3. 3
    Consumer boycott 3+ weeks; material sales impact
  4. 4
    Target public framing characterized rollback as 'refinement,' incomplete harm acknowledgment
  5. 5
    Rotation backfill assessment — first formal assessment for Target

These findings arrive in your inbox every Monday. Free.

Source intelligence

Primary-source alerts from overnight scanning. Each alert is linked to original regulatory filings, court records, investigative reports, and international legal instruments.

Fortune 500 — Worker Safety: Amazon PDX9 Death

An Amazon worker died at the Troutdale, Oregon PDX9 facility on April 6, 2026. Management ordered other employees to continue loading trucks for over an hour as the body remained on the floor. Oregon OSHA's 'non-work related' classification applies only to reporting requirements, not to workplace causation. Employees cite heat conditions caused by newly installed soundproof curtains. Coverage spread April 13-16 across multiple major outlets. This is the most serious worker safety incident flagged in the current scan window and comes on top of Amazon's first-ever Teamsters bargaining order (April 3) and the NLRB Palmdale contractor settlement (April 13).

amazon
  1. 1.techcrunch.com
  2. 2.novaramedia.com

Countries — Active Atrocity and Repression Events

Three significant April 2026 country-level events not yet captured in prior assessments: (1) Russia declared Memorial 'extremist' (April 9) and its youngest political prisoner (now 17) faces new charges (April 15) — political prisoner count now 1,217+. (2) South Sudan: OHCHR issued two April 2026 emergency alerts — civilian protection crisis and child trafficking crisis driven by SSPDF-SPLA-IO fighting; UNSC UNMISS mandate expires April 30. (3) Ukraine Easter ceasefire collapsed on Day 1 (April 11-12) — both sides alleged hundreds of violations; Russia refused extension. Iran's April 13 EU sanctions renewal and continued executions of January protesters is also actively developing.

russiasouth-sudanukraineiran
  1. 1.hrw.org
  2. 2.ohchr.org
  3. 3.aljazeera.com

Healthcare — UnitedHealth Group Q1 2026 Earnings Watch (April 21)

UnitedHealth Group Q1 2026 earnings release is set for April 21. This will be the first earnings call since CEO Stephen Hemsley returned after Brian Thompson's murder and the DOJ criminal Medicare fraud investigation became public. Full-year 2025 earnings fell 41%. Analysts expect Q1 2026 to be a pivotal disclosure event for the DOJ probe. The prior authorization reforms pledged after the December 2024 CEO shooting are in force as of January 2026 but are narrowly scoped. Multiple health insurer CEOs faced congressional grilling in January 2026 over claims denials. The April 21 earnings call is a major watch event for compassion-relevant governance disclosure.

unitedhealth-groupcignacvs-health
  1. 1.247wallst.com
  2. 2.insurancenewsnet.com
  3. 3.nbcnews.com

AI Labs — Illinois AI Liability Bills Committee Deadline April 24

The Illinois Senate committee deadline for SB 3444 (OpenAI-backed liability shield) and SB 3261 (Anthropic-backed strict safety/audit bill) is April 24, 2026. Senator Cunningham acknowledged on April 14 that SB 3444's sweeping liability relief provisions are 'highly unlikely' to survive intact, suggesting Anthropic's public opposition and the Gizmodo-covered 'OpenAI-Anthropic cold war in Illinois' may have shifted the legislative outcome. The April 24 vote outcome is a direct governance signal for both OpenAI and Anthropic's accountability posture scores.

openaianthropic
  1. 1.trackbill.com
  2. 2.gizmodo.com
  3. 3.transparencycoalition.ai

Get the full benchmark report

Daily briefings surface headline findings. Full benchmark reports include complete methodology documentation, all 40 subdimension scores, full evidence trails, certified assessments, and sector-level analysis packages.

Scores confirmed

Entities where research found published scores remain accurate. Confirmations are documented evidence, not silence.

EntityIndexBandPublishedAssessedDeltaDateFinding
AmazonFortune 500developing21.618.4-3.2PDX9 worker death and supervisor 'keep loading' order set Critical-band empathy floor; delta below 5-pt threshold.
South SudanCountriescritical03.8+3.8First baseline assessment; 267k+ displaced in Jonglei in 2026; UNMISS mandate vote before April 30.
UkraineCountriesfunctional5047.5-2.5Easter ceasefire collapsed; Ukraine refused Donetsk ultimatum; accountability institutions continue under wartime stress.
RussiaCountriescritical02.5+2.5Memorial designated 'extremist' April 9; teen political prisoner re-charged at 17; 51% prisoner-count increase since end-2024.
UnitedHealth GroupFortune 500critical16.914.4-2.5DOJ criminal MA fraud probe active; 2.3–2.8M members being exited; April 21 earnings call pivotal watch event.
IsraelCountriescritical8.89.4+0.6Six-month Gaza ceasefire scorecard confirms aid blockade; 150k children needing acute malnutrition treatment.
North KoreaCountriescritical01.3+1.3HRW March 2026 nuclear-program/human-rights linkage; UN Special Rapporteur: no improvement, some degradation.
EthiopiaCountriescritical5.98.8+2.91,300+ deaths from denied medicine/food in Tigray past 3 years; OHCHR Türk February 2026 warning.
Alphabet (Google)Fortune 500functional51.640.6-11AdX ruling imminent; $218B advertiser mass arbitration April 13–14; reassess when ruling drops. Delta vs displayed composite reflects prior methodology drift.
xAI/GrokAi Labscritical2.22.5+0.3NAACP v. xAI April 14: 27 unpermitted gas turbines in majority-Black community; already at index floor.
FinlandCountriesexemplary10097.5-2.5Rotation baseline; NATO integration tested but social services intact; Sami minority gaps persist.
DenmarkCountriesexemplary10096.3-3.7Rotation baseline; 'Ghetto laws' tested by ECtHR; core welfare system intact.
TIAAFortune 500exemplary92.587.5-5Rotation baseline; raw-score mean (87.5) differs from displayed composite (92.5) due to prior methodology; no material events.
Hugging FaceAi Labsexemplary90.988.1-2.8Rotation baseline; open-science mission operationalized; scaling pressure on BND noted.

Key highlights

Editorial-level findings from the Apr 18 research cycle.

01

The pipeline's first upgrade proposal has been generated. New York City's Preliminary Citywide Racial Equity Plan (April 6, 2026) produces the first upgrade in the pipeline's four-night history: +7.9, medium confidence, Functional band maintained. After 59 assessments across four nights generating zero upgrades, the first positive proposal is structurally significant institutional governance action from the highest-ranked US city in the index. Medium confidence is appropriate given implementation is pending; this should not be approved without the Q3 2026 checkpoint mechanism in place.

02

Target's DEI rollback is the sharpest values-under-pressure signal in the Exemplary band to date. A -10.0 delta on a rotation backfill entity that was never previously assessed is a methodologically important finding: Target's published score of 92.5 was set before the January 2025 DEI rollback, which this assessment flags as a material I1/EQU/ACC event. The band holds (Exemplary), but the 10-point decline is the largest delta among Exemplary-band entities in the pipeline. The finding is actionable without being catastrophic — Target remains above 80.

03

Three simultaneous Critical-band atrocity signals require human attention before April 30. The UNMISS mandate expiry (South Sudan, April 30), the Iran executions and EU sanctions renewal, and UnitedHealth's April 21 earnings are all within 13 days. South Sudan's UNMISS mandate vote is the highest-urgency political deadline in the pipeline — a mandate lapse would remove the last international protection layer from an active conflict zone where 267,000+ people have been displaced in 2026 alone. The scanner should treat any Security Council UNMISS developments as top-priority events through April 30.

04

The Alphabet delta-11 gap is accumulating without a proposal. Alphabet was confirmed at 42.2 on April 17 and assessed tonight at 40.6 (delta -11.0 from the published 51.6). The assessor held the proposal pending the AdX ruling due to low confidence (0.5). The AdX ruling has been overdue since March 31; at April 21-25 the pipeline will be in its fifth consecutive week of monitoring this event. If the ruling does not land by April 25, consider a methodology-drift-based change proposal at the current assessed level (40.6) regardless of the ruling — the 11-point gap from published reflects a real change, not a pending event.

The weekly briefing on institutional compassion scores

Score changes, sector trends, and emerging risk signals from overnight research across 1,155 entities — every Monday. Free.

No spam. Unsubscribe anytime. Your email is never shared.

Sector intelligence

Analyst-level observations on patterns emerging across indexed sectors from the Apr 18 research cycle.

Fortune 500 — DEI Rollback as a Scorable Event

    Fortune 500 — Big Tech Legal Accountability Cluster

    • Meta: $375M + $6M verdicts; Section 230 shield eroding across two jurisdictions
    • Alphabet: AdX ruling imminent; $218B mass arbitration filed April 13-14; YouTube addiction verdict in CA
    • Amazon: PDX9 worker death confirmed at Critical boundary; NLRB bargaining order stands
    • Boeing: Whistleblower family settlement ($50K net) remains the most concrete accountability symbol in the pipeline

    Countries — Active Atrocity Cluster First Baselines

    • Russia (2.5): Memorial banned as extremist, 1,217 political prisoners, documented torture. Score reflects absence of any positive dimensions, not a true zero.
    • North Korea (1.3): Closest to the absolute floor; nuclear-human rights linkage confirmed by HRW; no functional accountability institutions exist.
    • South Sudan (3.8): Active conflict-driven displacement crisis; UNMISS is the only international protection mechanism; mandate expires April 30.

    AI Labs — Illinois SB 3444 as a Dividing Line

      US Cities — NYC as Benchmark Setter

        Emerging risks

        Forward-looking risk signals from the Apr 18 research cycle. These are not current findings — they are early warning flags.

        Risk

        UNMISS mandate expiry April 30 (South Sudan). If the Security Council does not renew the UNMISS mandate before April 30, the UN Mission's 17,000 peacekeepers and civilian protection capacity lose their authorization. South Sudan assessed at 3.8 — the only international protection infrastructure in an active conflict zone with 267,000+ displaced in 2026 alone. A mandate lapse would immediately qualify for a major downgrade proposal. Scanner should treat any Security Council vote or veto threat as top-priority through April 30.

        Risk

        UnitedHealth Group Q1 2026 earnings April 21. This is the first earnings call since CEO Hemsley's return and the DOJ criminal Medicare Advantage fraud investigation became public. UHG is currently at 16.9 published (14.4 assessed), within 3.1 points of the Developing/Critical boundary. If the April 21 call includes DOJ cooperation disclosures, scope expansions, or civil settlement announcements, a downgrade proposal crossing the 20-point boundary is plausible. Treat as high-sensitivity event.

        Risk

        Alphabet AdX antitrust remedy ruling, expected April 21-25. The ruling has been overdue since March 31 (Judge Brinkema's self-imposed deadline). When it lands, Alphabet will need immediate reassessment. The assessed composite tonight is 40.6 (vs. 51.6 published, a -11.0 gap). A forced AdX divestiture ruling would affect ACC and SYS dimensions and likely produce a proposal at the assessed level. The $218B advertiser mass arbitration filed April 13-14 adds a second parallel accountability track. This is the most consequential pending single event in the Fortune 500 index.

        Risk

        Illinois SB 3444 committee vote April 24. Two AI Labs entities in the assessment queue are directly affected: OpenAI (currently flagged for review) and Anthropic (monitored). A liability-shield passage would be a negative signal for OpenAI's ACC/INT dimensions and could trigger a change proposal from the existing flag. An outcome stripping the liability relief provisions would be neutral-to-positive. The scanner should retrieve the committee vote outcome immediately on April 24.

        Risk

        Amazon PDX9 trajectory. Assessed at 18.4 tonight (21.6 published, -3.2 delta). The PDX9 worker death and supervisor conduct are documented but the Oregon OSHA investigation is not resolved. If OSHA reclassifies the death as work-related (likely given the heat/ventilation evidence), or if NLRB enforcement of the bargaining order escalates, Amazon could cross the Critical/Developing boundary downward. Currently 1.6 points below the applied Critical-band threshold. Monitor Oregon OSHA investigation outcome.

        Research insights

        Analytical observations from the Apr 18 research cycle. These are assessor-level interpretations, not findings.

        Note

        The pipeline has now produced its first upgrade after 59 assessments across four nights. The ratio of downgrade proposals to upgrade proposals across the pipeline's history is currently 18:1 (approximately 18 downgrade proposals, 1 upgrade). This is not a methodology artifact — the priority queue is explicitly weighted toward entities with recent negative news. But the near-absence of upgrade proposals is itself a finding: institutional compassion among major global entities is declining or stagnant in the 2025-2026 period. The NYC upgrade is the exception that makes this pattern visible.

        Note

        The Exemplary band has now been tested from both ends. Finland (97.5), Denmark (96.3), and TIAA (87.5) hold their Exemplary scores within tolerance on rotation backfill. Target (92.5 -> 82.5) is the first Exemplary-band entity to receive a downgrade proposal, and it remains in the Exemplary band. This band is not systematically inflated — the Nordic countries are substantiated. What Target demonstrates is that even well-performing entities can have specific dimension failures (EQU, ACC, INT) large enough to produce a material delta. The Exemplary-band confirmation rate (3 of 4 entities confirmed, 1 proposed downgrade) is the highest of any band in the pipeline's history.

        Note

        The accountability lag pattern is now visible across three industries simultaneously. Ford (18-month recall delay), Meta (platform design choices from 2012-2019 generating 2026 verdicts), and UnitedHealth (Medicare billing patterns from 2010-2023 generating a 2026 DOJ probe) all share the same structural dynamic: harm events are documented in real time, accountability events follow on a multi-year delay. The benchmark's methodology, which scores both the harm and the accountability response, captures this lag. The implication: entities currently in the Functional and Developing bands are likely to face downward pressure as prior-period harm events produce present-period accountability outcomes.

        Note

        Published scores for first-time Exemplary and Established entities are generally more reliable than for Critical-band entities. The pattern across 59 assessments: Critical-band entities assessed against primary sources (litigation, enforcement, OHCHR reports) receive large negative deltas when previously scored on ESG self-reporting (J&J -20.9, Mistral -29.5, CVS -18.7, OpenAI -20.2). High-scoring entities confirmed on rotation show small deltas (Finland -2.5, Denmark -3.7, TIAA -5.0, Iceland -1.6, Sweden -3.1). The systematic overstatement is concentrated in the Critical and Developing bands, where ESG infrastructure scores were credited without adequately weighting harm events. This is the most important calibration insight for the benchmark's published methodology.

        Want the complete picture?

        Full benchmark reports include all 40 subdimension scores, complete evidence trails, and methodology documentation for every assessed entity.