Study: Young users (9 to 18Y) ditch Google for AI, with unknown consequences
Young people – and many adults – are increasingly turning to AI chatbots instead of Google Search for answers, attracted by conversational queries, ad-free interfaces, and more direct, synthesized responses. Commenters argue that traditional search results have been “enshittified” by ads, SEO spam, and degraded ranking algorithms, making raw web search frustrating or unreliable. At the same time, they worry about new risks: manipulation and hidden advertising in AI outputs, uneven access to higher-quality paid models, weakened critical thinking and source-checking habits, and the broader shift from an open web of documents to AI-mediated information.
Shift from Google to AI tools
- Many commenters say they now default to LLMs/AI assistants for most “nontrivial” queries, with Google as a fallback.
- People report AI is faster, feels more conversational, and better handles natural-language, “vague” questions than traditional search.
- Some already get more traffic or customers from AI recommendations than from Google search.
Why many feel Google search degraded
- Complaints: SEO spam, cookie/pop-up clutter, intrusive ads, paywalled or low-value pages, and “AI-written SEO fluff.”
- Several argue Google deliberately degraded search quality (e.g., to increase time-on-site and ad revenue).
- AMP, AI overviews, and privileging ad ecosystems are seen as part of an “enshittification” arc.
Perceived strengths of LLM-based search
- Single coherent answer plus follow‑up questions is considered a major UX win.
- Good for exploratory questions and for users who never learned “how to search” with keywords/operators.
- Some like hybrid use: AI to explain/summarize, then verify via linked documentation or sources.
Reliability, manipulation, and ads
- Many worry users over-trust AI, rarely checking sources or noticing hallucinations.
- Examples: models inventing facts or misrepresenting referenced links; Google’s AI drawing opposite conclusions from its own citations.
- Concerns about corporate or state actors shaping AI outputs, “AI-SEO,” covert ads via logit biasing, and future ad-heavy LLMs.
- Some note this is analogous to long‑standing issues with search-engine ranking and media manipulation.
Effects on learning and critical thinking
- Split views: some think AI use can support critical thinking via back‑and‑forth questioning; others fear “spoon‑fed” answers erode skills and comparison of sources.
- Worries that children and adults alike will treat AI as infallible authority.
- Discussion that “critical thinking” is often poorly defined and already stressed in test-driven school systems.
Broader context & energy concerns
- Parallels drawn to earlier panics about students using Google, the web, or Wikipedia instead of libraries.
- Some see AI as just the next access layer to internet content.
- One thread raises environmental impact of GPU-heavy inference; replies argue batching and potential green energy lessen the concern, while cognitive impact is seen as more critical.