Anxiety Around New Technologies
Since the mid-15th century, every new technology impacting the written word has been met with anxiety. Each innovation—from the printing press to large language models (LLMs) like ChatGPT—has faced criticism for potentially diluting the link between thought and word.
Impact of Large Language Models
- LLMs, unlike other tools, can "think" for the user, raising concerns about their impact on knowledge production.
- The widespread use of LLMs, especially among students and researchers, has heightened these anxieties.
MIT Study on Cognitive Debt
- A recent MIT study explored the cognitive effects of using ChatGPT for writing.
- Participants were divided into three groups: those writing unaided, those using search engines, and those assisted by ChatGPT.
- Neural activity was monitored over four months, revealing lower cognitive engagement in the ChatGPT group.
- These users also showed less ownership and memory of their work.
Concerns and Optimism
- Fears persist that AI might replace intellectual labor, akin to automation in manufacturing.
- However, both the MIT study and AI evangelists like Sam Altman might overestimate AI's consequences.
Effective Use of AI
- While reliance on LLMs can have adverse cognitive effects, there's potential for effective use.
- Good writing involves analysis and expression, which can benefit from new tools.
Integration into Learning
- The stage at which AI is introduced into learning is crucial. For example, students learn long division despite calculators being available.
- LLMs can assist with language tasks like grammar correction without replacing authorship.
Research and Social Media Lessons
The rapid advancement of AI means research into its effects is lagging. Lessons from social media underscore the need to address this gap.