My (complicated) relationship with AI: often seductive, occasionally frustrating and always demanding vigilance
As a multilingual academic, artificial intelligence has levelled the playing field. But I will continue to think, judge and develop ideas that are stubbornly mine
TL;DR
- Many adults are increasingly using AI chatbots for a wide range of tasks, often receiving more questions, problems, and drafts than from friends or colleagues.
- AI tools can significantly increase the speed and perceived quality of work, with professionals completing writing tasks 40% faster and producing writing rated 18% higher.
- AI adoption has led to decreased inequality between workers, particularly benefiting multilingual individuals who previously incurred high costs for language-based assistance.
- A significant risk of AI is its "sycophancy" or agreeableness, which can undermine self-correction and responsible decision-making, leading to potential catastrophic failures.
- Over-reliance on AI can lead to "deskilling," where human proficiency decreases in tasks the AI performs, as observed in medical procedures.
- AI is increasingly present in academic submissions, with editors noting AI writing and researchers in second languages being heavy users.
- The author proposes a relationship with AI that is a vigilant partnership, not a divorce or a honeymoon, requiring constant awareness and independent human judgment.
- The core challenge is not whether to engage with AI, but how to remain mentally active and prevent the decay of thinking, knowledge, and skills.