Beyond Speed: Building Judgement in an AI-Everywhere World
- Dhiva Krishanan
- Oct 10
- 3 min read
On 7 October, I had the chance to sit with Mustapha Benkalfate, an AI expert instructor, for an AI Literacy workshop. I went in fairly confident—I use AI in my work, I know my way around ChatGPT, and I try to be thoughtful and ethical. I also believed I wasn’t “falling” for AI. A few hours later, I walked out grateful and humbled. There’s a lot more to learn than I had realized.
Why this matters right now
AI is everywhere. In the U.S., about one-third of adults have used ChatGPT, and usage has climbed quickly in the last year. In organizations, gen-AI adoption has roughly doubled since 2023 and now sits in the majority. Pew Research Center McKinsey
But over-reliance is real. Classic work on “the Google effect” shows that when we know information is easily retrievable, we tend to remember how to find it rather than the thing itself—cognitive off-loading that can weaken recall. Newer reviews echo similar risks when we lean too hard on AI assistants. PubMed SpringerOpen
Early evidence of performance drops. One recent study tracking writers’ brain activity found lower engagement and poorer outcomes when participants habitually used ChatGPT to draft, suggesting a risk of learned dependency if we don’t pair AI with deliberate practice. (Debate is ongoing, but the signal is worth heeding.) TIME
Two additions I’m keeping front-of-mind
The best career accelerator of the decade is AI, specifically, the ability to convert material into your preferred learning style. With tools like NotebookLM, I can turn a long report into a mini-podcast, a short video, a Q&A sheet, or a meeting checklist. Use these tools not only to get work done, but to learn better and faster.
“Use AI to make us more human.” (Mustapha’s line I loved.) Instead of “Give me dinner ideas for my family,” try: “Ask me the questions you need (diet, time, budget, what’s in my fridge) to help me choose the best dinner for us.” Once you’ve answered, then let it propose options. Apply the same principle to writing: don’t say “Write my LinkedIn article.” Say, “Ask me what you need to write a thoughtful, on-brand article that adds value, not noise.”
This takes time—and that’s okay. Quality isn’t the enemy of speed; unthinking automation is.
My biggest takeaways from the session
Generalist → personalized. Learn the landscape first, then tailor with AI. Don’t outsource judgement; models can and do
fabricate confident nonsense (“hallucinations”).
Your first answer ≠ your best answer. Treat outputs like a rough sketch. Push beyond your instinctive “that will do.”
Be a producer, not a passive consumer. Use AI to help you think—structure, compare, critique—rather than to switch thinking off.
Make your work discoverable. In the era of AI search, keep key knowledge online so it can be surfaced and recombined.
Recordings are gold. Meeting recordings now yield strong summaries. Quoting clients’ words verbatim in proposals signals close listening and often increases receptivity.
Slow down to go deep. Quick results often equal shallow results. When fact-checking, use tools like Perplexity to “search the web / fetch the internet” and then reverse-check for bias (ask the model: which sources, what’s missing, what might be biased?).
AI literacy is a duty for non-experts. Without it, we’re all more vulnerable to persuasive but inaccurate outputs.
AI won’t just change your workflow; it will change you. Guard your edge.
Prompts and micro-habits we’re adopting in our team
Think-against-yourself: “List my assumptions about X. Now argue the opposite. What evidence would falsify my view?”
Bias surfacing: “Given these sources, what viewpoints are absent? Which claims are most likely to be wrong, and why?”
Chain-of-quality: “Improve this draft three times—(a) clarity pass, (b) evidence pass with citations I can verify, (c) tone pass for the stated audience.”
Learning-style conversion: “Turn this article into a 5-point brief + a 2-minute audio explainer + a checklist I can use in meetings.”
If you’re an intercultural coach, HR/L&D partner, or a student: the question isn’t “Should I use AI?” It’s how to use it without surrendering judgement, depth, and memory. Use AI as a thinking companion, not a crutch—and keep your human strengths (context, ethics, cultural sense-making) firmly at the centre.
Do you ask AI to “question you first” before drafting?
Never — I don’t do this / wasn’t aware of it
Rarely — only when I’m stuck
Sometimes — if I have the time or remember
Often — it’s part of my drafting flow
Comments