Publication date: 09/23/2025
1. Overview – “What’s this all about?”
| Point | Info | |-------|------| | Author | Holger Schmidt | | Medium | Frankfurter Allgemeine Zeitung (FAZ+), Digital Economy | | Link | https://www.faz.net/pro/digitalwirtschaft/kuenstliche-intelligenz/ueberraschende-ergebnisse-so-nutzen-menschen-chatgpt-wirklich-110687940.html | | Publication date | 09/23/2025 |: 23.11.2025 | Read on | (today) | | Stated reading time | 5 min | | Realistic reading time (ignoring pop-ups) | approx. 4–6 min |
2. Summary – “I got it, now you do too”
Topic: OpenAI reveals for the first time in detail how humanity really uses ChatGPT in everyday life and at work.
- By the end of July 2025 ChatGPT reaches 700 million weekly active users and processes 2.5 billion messages per day.
- More than half of all work-related prompts in management & business revolve around text creation and editing.
- Main use cases: advisor role (ideas, decisions) and editorial tasks (writing, revising).
- The speed of adoption is considered historically unique—no other tech product has grown this fast.
- Communication and search habits are visibly shifting from the browser to dialogue with large language models.
- OpenAI speaks of “surprising results,” but provided only excerpts of the data—the complete dataset remains internal for now.
[⚠️ To be verified] - Monetization: FAZ+ seizes the opportunity to promote two subscription specials—synergy or banner blindness, who knows.
3. Opportunities & Risks – “It’s complicated”
Opportunities
- Efficiency gains: routine texts in minutes instead of hours.
- Democratization of knowledge: everyone has their “mini-advisor” in their pocket.
- Innovation boost: faster prototyping of ideas, products, campaigns.
Risks
- Dependency: if the bot goes on strike, the PowerPoint stays blank.
- Quality control: hallucinations can become expensive.
- Data sovereignty: companies hand confidential info to a black box.
[⚠️ To be verified]
4. A Look Ahead – “What else could happen?”
- Short term (1 year): More feature packs, integration into office suites, even more user records.
- Mid term (5 years): Specialized corporate agents, regulatory guardrails, new job profiles (“prompt strategist”).
- Long term (10–20 years): Dialogue-based interaction becomes the standard UI; whoever masters language rules the market—or we end up in a Netflix episode full of chatbots.
5. Fact Check – “Is this even true?”
| Statement | Plausible? | Comment | |-----------|-----------|---------| | 700 million weekly active users | ✔️ | Fits external traffic estimates. | | 2.5 billion messages/day | ✔️ | Reliable as long as OpenAI doesn’t report otherwise. | | “More than half” of business prompts are text-related | ✔️ | Sounds realistic, but should be defined more precisely. | | “Historically unique spread” | ⚠️ To be verified | Need comparison with smartphone apps like TikTok. | | Open data situation | ❌ | OpenAI shares only excerpts; full transparency is missing. |
6. Short Conclusion
ChatGPT is no longer just a nerdy toy but a mass-market tool for consulting and text work. The rapid adoption shows how quickly workflows shift, while questions of quality, responsibility, and data protection remain open. Companies should seize opportunities but have governance, transparency, and an emergency plan ready.
7. Three Critical Questions
- With ever-growing dependencies on OpenAI, is our freedom being quietly tightened?
- Responsibility: Who is liable if a CEO makes bad decisions based on hallucinated data?
- Does widespread use foster real innovation—or just high-speed copy-paste content without transparency on how the results actually arise?