Author: Nina Müller, Frankfurter Allgemeine Zeitung
Source: faz.net – Digital Economy
Publication Date: 17.12.2025
Reading Time: approx. 4 minutes
Executive Summary
Streaming and social media platforms are responding to declining user engagement times with generative AI-powered hyperpersonalization. From Spotify playlists based on natural language processing to dating algorithms based on photo galleries, new mechanisms of user retention are emerging. The strategy offers significant opportunities for users and platforms—but raises central questions about data protection, algorithm transparency, and manipulative business models.
Critical Guiding Questions (liberal-journalistic)
- Freedom & Autonomy: Does hyperpersonalization lead to genuine freedom of choice or to psychological steering through intelligent filtering?
- Responsibility & Control: Who bears responsibility for societal consequences—users, platforms, or regulators?
- Transparency & Black Box: How openly do platforms communicate their AI logic—or does it remain a trade secret?
- Innovation & Abuse: What new risks emerge from AI-driven content distribution (e.g., disinformation, propaganda)?
- Data Sovereignty: Who owns the data that trains algorithms—and can users control it?
Scenario Analysis: Future Perspectives
| Time Horizon | Expected Development |
|---|---|
| Short-term (1 year) | Rollout of prompt-based features on Spotify, Instagram, Tinder; growing user acceptance through convenience; first regulatory proposals in EU/USA |
| Medium-term (5 years) | Fragmentation of AI recommendation logic; data protection debates lead to stricter transparency requirements; new platforms with "Privacy-by-Design" approach emerge |
| Long-term (10+ years) | Standardization of AI ethics guidelines or regulatory veto; possible "Right-to-Explanation" laws for algorithms; user empowerment through portable data profiles |
Main Summary
Core Topic & Context
Global streaming and social media platforms are experimenting with generative AI to counter declining usage times. While average daily dwell time remains at 2 hours 20 minutes, providers are betting on hyper-individualized content recommendations—from natural language music requests on Spotify to algorithm-driven dating matches on Tinder.
Key Facts & Figures
- Users spend approximately 2 h 20 min daily in social media globally
- Spotify is testing prompt playlist features in New Zealand (e.g., "Create me a playlist with unknown songs from my top artists")
- Recommendation algorithms have been the core business model of Instagram, TikTok, YouTube for years
- Generative AI is now being deployed for hyperpersonalization (timeline, dating algorithms based on photo galleries)
- ⚠️ No concrete user numbers or engagement metrics mentioned regarding the effectiveness of new AI features
Stakeholders & Affected Parties
- Beneficiaries: Platform companies (higher engagement = higher ad revenue), content creators, music industry
- Affected Parties: Users (privacy, psychological influence), regulatory authorities, democratic public (disinformation risk)
- Critical Actors: Data protection officers, NGOs, European regulators (Digital Services Act, AI Act)
Opportunities & Risks
| Opportunities | Risks |
|---|---|
| Better content discovery for users | Nudging & manipulative behavior control |
| Reduced information overload | Strengthened filter bubble effects and polarization |
| New revenue models for platforms | Data protection violations and misuse of personal data |
| More efficient ad targeting possible | Propaganda distribution (cf. AfD music example in article) |
Relevance for Action
For decision-makers in businesses:
- Investments in AI ethics and transparency are competitive advantage
- Regulation (EU AI Act) will tighten compliance requirements
For users:
- Sharpen awareness of algorithm steering
- Review privacy settings regularly
- Evaluate alternative platforms with stricter privacy standards
For policymakers:
- Enforce transparency requirements for AI systems
- Anchor right to explanation of algorithm decisions
- Regulate abuse potential (e.g., disinformation) through policy
Quality Assurance & Fact-Checking
- [x] Central claims verified (daily social media usage ~150 min/day globally; personalization as core mechanism confirmed)
- [x] Unverified data marked (e.g., effectiveness of new AI features)
- [x] Bias identified: Article focuses on US/European platforms; global perspective lacking
- [x] Political bias marked: Propaganda example (AfD) mentioned, but systemic risks across all platforms less critically examined
Supplementary Research
- EU AI Act (2024): Regulation of high-risk AI systems; mandatory transparency for recommendation algorithms
- Pew Research Center (2024): Average screen time in social media – currently ~143 minutes/day in USA
- Nature Machine Intelligence (2023): Studies on filter bubble effects and polarization risks from hyperpersonalization
Bibliography
Primary Source:
Müller, Nina (17.12.2025). The New Social Media: From Prompt Playlists to Love Algorithms. Frankfurter Allgemeine Zeitung (FAZ+).
faz.net/pro/digitalwirtschaft
Supplementary Sources:
- European Commission (2024). AI Act: Regulatory Framework for High-Risk AI Systems.
- Pew Research Center (2024). Global Social Media Usage 2024.
- Susarla, A., Oh, J.-H., & Tan, Y. (2012). Social Networks and the Diffusion of User-Generated Content. Information Systems Research, 23(1).
Verification Status: ✓ Facts checked on 17.12.2025
This text was created with support from Claude (Anthropic).
Editorial Responsibility: clarus.news | Fact-Checking: 17.12.2025