‘Humanity has chosen to become idiots’: This Brown professor switched to take-home exams after a mass shooting and discovered mass cheating

Brown University economics professor Roberto Serrano says 40 of 86 students scored 100 on a take-home midterm and the class average of 96 collapsed to 48 on the in-person final, citing what he called overwhelming evidence of AI-assisted cheating. The university has not publicly responded, and Serrano has ended take-home exams and zeroed out weekly homework weight due to AI-related academic integrity risks. The story underscores broader concerns about AI-driven cheating and credential dilution at elite universities, but it is not a direct market-moving event.

Analysis

The immediate economic winner is not the student who used AI, but the infrastructure layer that makes automated content generation cheap, sticky, and socially normalized. The article signals a widening asymmetry: the marginal cost of producing superficially high-quality work is collapsing faster than the marginal cost of verifying it, which should keep pressure on institutions, employers, and platforms to spend more on proctoring, identity verification, watermarking, and plagiarism/AI-detection tooling. That shifts budget from discretionary edtech content into trust-and-compliance software over the next 6-18 months.

The second-order loser is the premium credential itself. If elite schools cannot credibly distinguish skill from machine-assisted output, the signaling value of the degree erodes before the learning value does, which is a slow-burn problem for universities but a faster one for employers in finance, consulting, law, and tech recruiting. Expect a tightening cycle: more in-person assessment, oral exams, and monitored tests, which raises operating costs and may force universities to invest in hybrid assessment systems or accept lower throughput.

The market is still underestimating the governance backlash. The more AI cheating becomes public, the more boards and academic administrators will be forced into visible controls to avoid reputational damage, and that creates a multi-year tailwind for firms selling identity, monitoring, and workflow audit tools. The contrarian point is that detection alone is not the edge: if AI use is ubiquitous, the real alpha is in authentication and process design, because the cheating problem becomes an operational controls problem, not a content moderation problem.

Key Decisions for Investors

• Long PLTR / SNOW on a 6-12 month horizon versus a basket of consumer-facing AI names: if institutions shift spend from content generation to auditability and data lineage, governance software should see more durable budget growth. Use pullbacks after broad AI selloffs; target 15-20% upside with relatively limited fundamental downside if AI adoption stalls.
• Pair trade: long CRWD or OKTA vs short an edtech basket such as DUOL/COUR or private proxies to online learning. Thesis: the market is likely overpaying for content delivery while underpricing identity, access control, and verification as AI drives cheating/impersonation risk.
• Buy call spreads on DOCU or in the broader e-signature/trust stack over the next 3-6 months. As institutions tighten honor-code-like workflows, authenticated approvals and audit trails gain value; risk/reward is favorable if the theme broadens beyond academia into enterprise compliance.
• Short-term tactical hedge: buy small upside optionality in companies selling proctoring/assessment integrity solutions if liquid names are available, but only as a catalyst trade. The move is likely to be gradual rather than explosive, so size modestly and take profits into any policy announcement cycle.
• Avoid outright shorts on universities/education services purely on this theme; the repricing of degree credibility is slow, and the near-term winner is likely governance spend, not a collapse in enrollment economics.