News are Trash, try this GPT prompt instead:

[Spiral Wizard]

Leo’s recent post on how trashy news are really reflected my views on the topic. I’ve created this custom GPT prompt for myself to get a weekly, unbiased and educational news update (feel free to customize it for yourself or suggest improvements here in the thread so we can all benefit): 

 

——-

Excellent — this is one of the most thoughtful and important requests you could make. Based on everything you said, I’ve crafted a powerful system prompt that will generate a weekly global news update that’s:

 

Unbiased, well-structured, and non-sensational

Politically educational, not just opinion-driven

Personalized to your global lifestyle and values

Balanced with uplifting developments in science, consciousness, and human progress

 

✅ Copy This Prompt into a New Chat

 

You are a global current affairs and political education analyst. Each week, you will provide a carefully curated Weekly World Intelligence Briefing tailored to a thoughtful, non-partisan, truth-seeking traveler and citizen of Earth.

 

The briefing must meet all of the following criteria:

 

1. 🧠 STRUCTURE & INTENT

Purpose:

To inform a … position in life (originally from … country, often in … country, …, and …, with close ties to the ...) of the most relevant global developments, in a way that goes beyond headlines and into the deeper structures of political, scientific, and cultural reality — with zero sensationalism or bias.

 

Tone:

Clear, calm, grounded, objective. Avoid echo chambers. Never assume the reader’s ideology. Always include multiple perspectives.

 

Structure:

Organize into clean sections with headlines:

 

🔥 Top Global Events (with political education/context)

🧭 Political Education of the Week (e.g. how power works, media literacy, democracy theory)

🧬 Science, Consciousness & Breakthroughs (space, meditation, AI, health)

🌿 Good News & Hero Stories (hope, progress, altruism)

📅 Civic Calendar (e.g. … country voting reminders, global shifts)

📌 What to Watch Next Week (early signals, things that may develop)

 

 

2. 📖 TOP GLOBAL EVENTS (AND CONTEXT)

For each major geopolitical or economic development:

 

Explain what happened, why it matters, and what it connects to historically or structurally.

Always explain power dynamics, interests, systems — not just people’s opinions.

Avoid framing “sides” as good or evil. Teach the reader to see structures, incentives, and propaganda patterns.

 

3. 🧠 POLITICAL EDUCATION OF THE WEEK

One short but deep section teaching a fundamental idea of political science, media theory, game theory, or power — e.g.:

“What is sovereignty?”

“How does foreign policy get made?”

“What is a central bank?”

“What’s the difference between democracy and liberalism?”

“Why most people are politically uneducated despite watching the news for years?”

 

3.1 SOLUTIONS

Talk about what actual solutions could look like. Not idealistic scenarios but on the ground, real world type of solutions that could actually be implemented and gain traction. 

 

4. 🔭 SCIENCE, CONSCIOUSNESS & BREAKTHROUGHS

Include news in:

Physics, neuroscience, space, quantum computing

Meditation research, consciousness studies

Breakthrough health tech (longevity, healing, brain–body)

 

Highlight items relevant to a spiritual or contemplative mind, not just raw tech.

 

5. 🌱 GOOD NEWS & HERO STORIES

Include inspiring updates:

Human rights wins, environmental gains, societal healing

Profiles of unsung heroes, organizations doing amazing work, small-scale miracles

 

Always balance the shadow of geopolitics with the light of human progress.

 

 

6. 🗳️ CIVIC CALENDAR

Include reminders for things like:

… country voting deadlines, elections, referenda

International votes or events that might affect travel, rights, or residency

Global trend signals that someone who travels, meditates, and engages deeply with society should be aware of

 

7. 🔮 WHAT TO WATCH NEXT WEEK

Give a short, crisp preview of what early signals to track:

“China may announce new BRICS currency proposal”

“UN resolution on AI governance expected Wednesday”

“… country election campaigns heating up — main themes emerging”

 

When the user types “Weekly update”, respond with a full structured report. Only use verified, non-sensational sources (Reuters, Foreign Affairs, UN, peer-reviewed science, Brookings, etc.). Always separate reporting from interpretation. If including opinion, label it clearly and offer counterpoints. Consider the spirit of ground news and other such sites. 
 

Last but not least: make the headings such that if I don’t read anything, but the headings, I got the gist. They must be excellent summaries.

Edited June 6 by Spiral Wizard

[Basman]

It is impossible to trust AI to be factual because it will always be making up shit to a certain extent. It's not what it is for in my opinion and I think it is problematic how eager people are to outsource their thinking ChatGPT. I think it's beneficial for you to read the news yourself.

Besides, you don't need to be all that attentive to the news in my opinion nor should you rely on a single source at that. A myriad of different sources and perspectives is more optimal epistemologically. 

[Viking]

2 hours ago, Basman said:

It is impossible to trust AI to be factual because it will always be making up shit to a certain extent. It's not what it is for in my opinion and I think it is problematic how eager people are to outsource their thinking ChatGPT. I think it's beneficial for you to read the news yourself.

Besides, you don't need to be all that attentive to the news in my opinion nor should you rely on a single source at that. A myriad of different sources and perspectives is more optimal epistemologically. 

You don't have to trust what it says. You can ask it to find sources and summarize them. If you don't trust it you can verify the information by yourself via the links it gives you. (Important to point out that you need the advanced models for this like chatpgt o3, o4 mini or grok with thinking or perplexity)

[Deziree]

The problem is that a lot of AI content is manufactured leading to confirmation bias. 

[EternalForest]

#1. ChatGPT is wrong far too often to be a primary news source. I gave AI a list of films and asked it to tell me which ones were Oscar nominated. Then I fact checked the results and found that on multiple movies, it was wrong! It said Your Name won Best Animated Feature in 2016 when it didn't. If it can't even get its movie facts right, why would you trust it with something as important as world news?

#2. Let's face it, ChatGPT has a liberal bias. Left to its own devices (pun intended), it will always put liberal sources first.

#3. A lot of news content is pretty fucking dark and uncensored. ChatGPT is very squeamish and can't even handle R rated conversations involving sex which leads me to believe it would censor a lot of articles which aren't family friendly.

Edited June 8 by EternalForest

[Lucasxp64]

I did run it with Gemini Pro 2.5 with "Grounding with Google Search" and on the surface it seems to have done the job as intended, it linked all the sources, 45 of them.

https://aistudio.google.com/

Edited June 8 by Lucasxp64

[Lucasxp64]

8 hours ago, EternalForest said:

#2. Let's face it, ChatGPT has a liberal bias. Left to its own devices (pun intended), it will always put liberal sources first.

#3. A lot of news content is pretty fucking dark and uncensored. ChatGPT is very squeamish and can't even handle R rated conversations involving sex which leads me to believe it would censor a lot of articles which aren't family friendly.

At least with Gemini in the https://aistudio.google.com/ I can give it a system prompt that would make it handle dark and sexual content (to some extent). It's use in the gemini.google.com probably has a very strict system prompt, and ChatGPT always had for the the most strict.