Intelligence Brief
The research, frameworks, and signal that separate operators from spectators. Read and summarized so you do not have to.
Published Monday, September 7 at 6:00 AM PT. 14 sources, 5 minute read.
What makes humans irreplaceable in the AI era.
HBR Staff
Why it matters. A Harvard field experiment had 228 experienced evaluators assess 48 innovation submissions under human-only and LLM-assisted conditions. The signal for SupaHumans: judgment is not a by-product of having more intelligence, it is the capability to interrogate evidence, retain context, and own the decision.
Erik Brynjolfsson, Bharat Chandar, and Ruyu Chen
Why it matters. Using ADP payroll data, Stanford finds workers aged 22 to 25 in highly AI-exposed occupations are about 19% below the employment level they would have reached relative to less-exposed peers, while experienced workers show no comparable gap. The emerging edge is not generic AI fluency, it is the tacit judgment and adaptive capacity that complement automation.
Sagar Goel, David Martin, and Charikleia Kaffe
Why it matters. BCG's survey of 70 C-suite and senior executives found half already see organizational de-skilling and more than 60% expect it to become material within three to five years. Judgment, problem framing, causal reasoning, creative thinking, and solution evaluation are capabilities organizations must deliberately practice and measure.
Why resumes, interviews, and hiring infrastructure are failing.
Marvin Starominski-Uehara
Why it matters. WEF cites Greenhouse's 2025 survey of 4,100 recruiters and hiring managers: 34% spend up to half their week filtering spam applications and 91% have caught candidates being dishonest. With ATS adoption rising from 26% of organizations in 2024 to 43% in 2025, the case for trace hiring is the case for evidence. Watch how a candidate collaborates and solves a real problem instead of trusting polished claims.
WorkplaceTech Spotlight
Why it matters. SHRM reports that companies lose an average of $17,000 per bad hire, managers spend 26% of their time coaching wrong hires, and white-sounding names received 9% more callbacks than Black-sounding names. Structured, evidence-based evaluation is both a fairness control and a margin-protection system.
Kara Dennison
Why it matters. Fabric analyzed 19,368 AI-powered interviews and flagged 38.5% for AI-cheating behavior, with the rate reaching 48% in technical roles. When AI can generate the resume, coach the interview, and feed real-time answers, hiring has to verify capability through work samples, practical scenarios, and evidence tied to the candidate's actual experience.
Rishi Bommasani, Sarah H. Bana, Kathleen A. Creel, Dan Jurafsky, and Percy Liang
Why it matters. Stanford's analysis followed 3.4 million people across 4 million applications, 1,700 postings, and 150 employers using one third-party AI vendor. Position-level analysis found adverse impact affecting 26% of Black applicants and 15% of Asian applicants, making auditability and additional proof of capability non-negotiable.
The behavioral science of who actually thrives.
Kevin Kruse, drawing on Travis Bradberry's assessment data
Why it matters. The cited analysis draws on more than 500,000 assessments and reports that EQ accounts for 58% of performance across job types, with 90% of top performers scoring high in it. As AI absorbs more technical execution, self-awareness, self-management, social awareness, and relationship management become observable performance levers, not personality decoration.
Deloitte US press room
Why it matters. Deloitte's survey of 1,394 workers found high-performing teams were more likely to use AI, but their strongest results were associated with curiosity, resilience, informed agility, connected teaming, and emotional and social intelligence. The predictor is not a static trait, it is adaptive behavior expressed in collaboration, trust, and learning.
Lauren Holzman via Business Wire
Why it matters. The report combines insights from nearly 700 leadership, HR, and L&D professionals with EQ data from more than 23,000 people. Adaptability, critical thinking, EQ, and communication rank alongside technology as the skills that matter most. It validates a practical standard: who stays steady, learns, and executes under uncertainty.
Where AI and human collaboration is actually heading.
Julian Berger et al.
Why it matters. A re-analysis of all 74 studies from a prior meta-analysis found that outcome feedback paired with AI explanations was associated with positive synergy, while explanations without feedback were associated with negative synergy. Collaboration improves when people learn to verify reliability. Adding AI is not an operating model.
Deloitte US press room
Why it matters. Deloitte's survey of 501 leaders plus 20 executive interviews found only 5% of organizations say their processes are highly prepared for AI agents, while 75% believe human collaboration with agents creates more value than agent-only automation. The real work is redesigning workflows, escalation paths, trust thresholds, and human decision rights.
Fabrizio Dell'Acqua et al.
Why it matters. In a preregistered randomized field experiment with 776 Procter & Gamble professionals, individuals using AI matched the performance of teams without AI, while AI-enabled teams were 9.2 percentage points more likely to produce a top-decile solution than the no-AI individual baseline. Integration with expertise and cross-functional context determines the upside.
Vaccaro et al.
Why it matters. A systematic review of 74 papers, 106 experiments, and 370 effect sizes found human and AI systems improved on human-only performance on average (Hedges' g = 0.64) but underperformed the better of human or AI alone (g = -0.23). Collaboration is not a magic property of adding a model. Calibration, design, and accountability decide whether it helps.
