–We need a fear-adjusted news metric

–To be clear, I’m not advocating for woke propaganda. I’m advocating for more stone-cold logic in boosting news that really matters to humans’ daily life, everyday mental focus, rather than triggering folks’ illogical (still kind-of-important) fears

–How AI could** help make real-world life less stressful

Opinion and research by Mathew Carr, ChatGPT

Feb. 20-March 7, 2026 — News-highlighting algorithms used by search engines, social media platforms and AI chatbots could and should be adjusted to better reflect real-world risks, instead of emotion-based fears, according to new research.

Here’s how it will happen (in my dreams). The best way to explain is by using vibrant recent examples.

While a flash flood in Gansu Province, China, killed around 10 people (probably many more) during a severe rainfall event last summer, the global media attention it received was far less than the coverage given to recent violent incidents in Canada and the UK with similar or much-lower casualty counts.

This illustrates how news systems and social algorithms tend to elevate stories with immediate personal‑fear triggers (school violence, a 13-year-old alleged perpetrator, shootings) while treating climate‑related loss of life as background context — even though both types of events reflect real risk to human life.

https://share.upscrolled.com/en/post/d1cc5050-0eba-11f1-8080-800053d5bf64/

Illustrated chart of readership/interaction with of school violence news v climate violence news

How could a news system be tweaked so that coverage of a smaller-but-objectively significant climate event isn’t under-amplified compared to violent incidents?

This is exactly where the fear-adjusted news metric logic comes in. Let’s break it down carefully.

1. Why the climate story underperforms in the current system

Emotional hooks are weaker: School shootings trigger immediate fear (“could happen to my child”), whereas a minor climate disaster feels distant or abstract.

Algorithmic reward bias: Social feeds optimize for clicks, comments, shares — which favor stories that trigger fear and outrage.

Story context is complex: Climate events require explanation about causality, broader risks, or long-term impact, which doesn’t fit the “instant shock” formula.

Result: a story like the China flash floods gets lower height in your chart, even though the real human impact is comparable.

2. How to tweak the system / algorithm

The goal: increase coverage for objectively significant but low-emotion events, without resorting to censorship.

A) Adjust the reward function

Currently:

Reward = f(engagement)

Engagement = clicks, shares, likes, comments

Emotional amplification = implicit

Proposed tweak:

Reward = \alpha \times Engagement + \beta \times Objective\_Severity

Objective Severity = metric of real harm, fatalities, injuries, displacement

α and β = weights; currently α ≫ β.

To boost climate coverage, increase β so that stories with genuine human impact get algorithmic amplification regardless of immediate emotional response.

B) Introduce “impact normalization”

Normalize across story categories: violence, natural disasters, health, climate.

Stories in under-covered categories get a multiplier to reach potential users.

For your chart, this would raise the China flash floods bar, potentially even above one violent story, reflecting its actual severity.

C) Reward contextual clarity

Algorithms could favor stories that provide contextual metrics, not just shocking imagery:

Number of people affected

Base rate of occurrence

Preventable or systemic risk

This means small climate events with 10–15 fatalities could compete with school violence stories in terms of coverage.

D) Decouple engagement from fear triggers

Explicitly reduce the weight of signals like “anger” or “shock” in feed ranking.

Increase weight of signals like “informative value,” “long-term risk,” “humanitarian relevance.”

Outcome: stories that are important but not fear-triggering are amplified.

3. How it would affect the chart

Result: the system aligns coverage with actual risk and harm, not just what drives clicks or panic.

4. Optional fine-tuning

Apply category-specific multipliers: e.g., small-scale climate disasters get +1.2–1.5 to reach potential.

Incorporate temporal persistence: even if a story doesn’t go viral initially, maintain visibility over several days to let its real-world importance register.

Include editorial override signals: a human-in-the-loop can flag objectively important stories for extra amplification, reinforcing the metric.

Upshot

If we implement these tweaks:

The China flash floods story could be amplified as much or more than the violent incidents, reflecting real human impact.

Readers get a more rational picture of what’s urgent: violent incidents are still covered, but climate-related human harm is no longer buried.

The system incentivizes fact-based, harm-proportional coverage instead of fear-driven amplification.

(* and ** : changed from “will” to “could” after further research)