Could AI Threaten the Growth of New Knowledge?

As AI tools like ChatGPT and Google Gemini become go-to sources for information, a provocative question arises: Will these systems stop us from producing new knowledge? In his latest commentary, Greg Ip warns that while AI is excellent at retrieving and synthesizing existing information, it doesn’t contribute fresh ideas — and overreliance may erode our incentive to create. 

Synthesis vs. Creation

Ip notes that AI models are trained on massive corpora of human‐generated text, and when asked for answers, they remix those inputs rather than invent. In contrast, human researchers and writers often explore novel angles motivated by curiosity, career, or impact. If AI dominates the “answering game,” the rewards for original thinking could diminish. 

Signs of Human Contribution Declining

The piece points to real data trends:

• Stack Overflow, once a vibrant Q&A hub for developers, has seen a dramatic drop in questions (over 90% in some languages) since AI tools gained popularity. As fewer developers post questions or answers, the knowledge base that future AI systems rely on is weakened. 
• Wikipedia pages similar to AI‐generated content have experienced declines in both views and edits — hinting that human contributors may be stepping back when AI seems able to supply what’s needed. 
• Web publishers have also suffered as AI surfaces content directly to users, reducing click-throughs and referral traffic — making it harder for creators to sustain their work. 

Risk of “Model Collapse”

One looming concern is recursive training: AI models increasingly learn from AI‐generated output, not original human work. Over time, this could cause a feedback loop where quality degrades (“model collapse”) because the foundation of knowledge becomes overly circular. 

Ip draws a parallel to the rise of passive investing: just as index funds ride on others’ research rather than doing their own, widespread AI reliance might erode the incentives for active knowledge creation. 

Can Creativity Survive the Shift?

Not all is doom and gloom. Some optimists argue that AI might free humans from rote tasks and let them focus more on creative, interdisciplinary, or speculative work. Economist Joshua Gans, for instance, suggests that as long as novel insight has value, people will still generate new ideas. 

But Ip sounds a cautionary note: what if we lose the habit of learning itself? Studies cited in the article showed that individuals using AI to write essays engaged less with memory and executive brain regions than those writing independently — raising fears that over time, our capacity for critical thinking could weaken. 

Takeaways

• AI is a powerful aggregator and synthesizer — but its role in creating new knowledge is limited.
• Human incentives to explore, innovate, and contribute might atrophy if AI dominates how we access information.
• The decline in participation on platforms like Stack Overflow and Wikipedia could be early warning signs.
• Maintaining a balance — using AI as a tool, not a replacement — may be crucial to preserving intellectual growth.