Alarming Study Finds That Most People Just Do What ChatGPT Tells Them, Even If It's Totally Wrong

45%: a BBC study found leading AI chatbots give incorrect answers 45% of the time. In UPenn experiments, over 50% of participants used ChatGPT voluntarily, following correct AI advice 92.7% of the time and following faulty recommendations 79.8% of the time, demonstrating strong user reliance or 'cognitive surrender.' The findings signal growing risks of diminished critical thinking and potential reputational, regulatory, and operational implications as AI becomes more deeply integrated into daily decision-making.

Analysis

Behavioral research showing widespread “cognitive surrender” is a demand-side shock with asymmetric implications — it simultaneously raises monetization optionality for trusted incumbents and increases liability/regulatory risk. Over months to a few years, large platforms that can credibly offer verification, provenance layers, and enterprise-grade guardrails will be able to charge an “AI safety” premium (subscription or ad CPM uplift) while smaller players will struggle to underwrite the compliance cost.

For Google specifically, the second-order channel to watch is ad attribution and search intent drift: if end-users increasingly accept AI-provided answers without clicking through, paid search CTRs and keyword marketplaces could shrink 3–7% in the near term, pressuring CPMs — but Google can blunt that by embedding paid “verified answers” or shifting more queries into monetizable surfaces. Regulatory and litigation tail risk is real and front-loaded: a few high-profile hallucinations could trigger consumer-protection inquiries and a cycle of forced transparency rules within 6–24 months, increasing short-term compliance costs (hundreds of millions to low billions) but favouring scale players who absorb them.

The contrarian angle is that what many see as an existential “loss of thinking” for consumers is actually a demand accelerator for trustable, auditable AI layers; market panic would hurt smaller, unprofitable AI challengers far more than Google. Net: modest short-term reputational volatility is likely, but structurally the platform that sells verifiable, monetizable safety wins; that asymmetry favors owning the incumbent selectively while hedging headline risk.

Key Decisions for Investors

• Core exposure: Buy GOOGL (or a 6–12 month bull-call spread) sized 1–2% of portfolio to express long-term monetization of paid ‘verified AI’ features and enterprise guardrails. Target +25–40% upside in 12 months versus a 12–15% downside stop — favorable asymmetric payoff if incumbents capture the safety premium.
• Near-term hedge: Buy a 3-month 8–12% OTM put (or put spread) on GOOG equal to ~25–33% of the GOOGL long notional to protect against a headline-driven 10–20% sell-off. Expect cost ~1–2% of notional; this creates a ~3:1 protection ratio versus realized downside scenarios.
• Tactical volatility capture: If shares gap down >7% on adverse headlines, sell covered calls on 25–50% of the existing GOOGL position to monetize elevated IV while using proceeds to add to puts or increase long exposure — target to recycle ~50% of collected premium into additional long exposure at lower cost.