You’re mid-explanation of a discounted cash flow model when the interviewer suddenly asks how many ping-pong balls fit in an airplane. It feels like a non sequitur — until you understand what’s actually happened to interviews in the last two years, and why that jarring shift in question type is now more deliberate than ever.
What’s Actually Changed: AI Cheating Tools Are Everywhere
Interview cheating has moved well past “someone Googled the answer beforehand.” A 2026 analysis of nearly 19,400 interviews found dedicated cheating tools like Cluely and Interview Coder account for close to half of detected cheating cases, with candidates using ChatGPT or Gemini’s voice mode for another sizable share — a phone or second screen quietly listening to the interviewer’s question and feeding back a polished answer through an earpiece or hidden overlay. Roughly two-thirds of hiring managers now say they’re specifically worried about candidates using generative AI to gain an unfair edge during recruitment assessments.
This isn’t a hypothetical problem for finance recruiters. It’s already visible enough that major firms are addressing it directly.
Why Goldman Sachs Is Pushing Back
Goldman Sachs recently warned its campus recruits against using ChatGPT during interviews, telling students to study the firm instead of leaning on AI tools — and reserving the right to disqualify candidates caught doing it. Amazon has issued similar guidance to its own recruiters, stating explicitly that generative AI tools give candidates an “unfair advantage” that masks a genuine read on their actual capabilities. Goldman uses the video interviewing platform HireVue for pre-assessment screening, part of a broader industry shift toward interview formats that are harder to game with a hidden screen.
That shift matters directly for brain teasers, because it turns out they were never really about the puzzle in the first place.
Why Brain Teasers Are Harder to Fake Than Technical Questions
Scripted technical questions have a predictable structure, which is exactly what makes them vulnerable — an AI model trained on the same textbooks interviewers pull rubrics from can generate a textbook-perfect answer in seconds. Brain teasers work differently. They require a candidate to think out loud, adjust assumptions in real time, and visibly work through ambiguity while the interviewer watches — which is far harder to fake convincingly than reading a polished answer off a hidden screen. AI-cheating detection research has found that tools reading back AI-generated responses tend to create a telltale delay between question and answer; a live, improvised brain teaser response doesn’t have that clean a script to hide behind in the first place.
| Interview format | How easy to cheat with AI | Why |
|---|---|---|
| Take-home technical test | Very easy | Unsupervised, unlimited time to consult AI tools |
| Scripted technical Q&A (live) | Moderate | Predictable structure matches AI training data closely |
| Live brain teaser / lateral thinking puzzle | Difficult | Requires visible, improvised reasoning that’s hard to fake convincingly in real time |
| Asynchronous timed video response | Difficult | Countdown timers leave little room to consult outside tools |
That’s part of why brain teasers, far from being an outdated hazing ritual, are becoming a more deliberate part of how banks separate real reasoning ability from a rehearsed or AI-assisted performance.
Understanding Brain Teasers in Investment Banking Interviews
Brain teasers investment banking interviews commonly feature puzzles or unconventional questions that test reasoning rather than financial knowledge — estimation, logic, probability, or lateral thinking, often with no formula or textbook solution. Interviewers care far less about a perfect answer and far more about how a candidate structures their thinking, handles ambiguity, and explains their logic aloud in real time — the exact thing an AI-generated answer skips past.
Common Categories
Brain teaser examples in banking interviews tend to fall into recognizable buckets: market-sizing questions (estimating the number of gas stations in a country), logic puzzles built around constrained scenarios, probability and math questions testing numerical intuition under pressure, and riddles that test lateral thinking and hidden assumptions. All of them share the same underlying goal — watching how a candidate thinks, not what they’ve already memorized.
Personal Experience: What Coaching Candidates Taught Me About This Shift
I’ve coached a few candidates through IB interview prep over the past year, and the ones who leaned too heavily on AI tools to draft “ideal answers” in advance consistently struggled the moment a live interviewer pushed back or changed the scenario mid-question. One candidate had memorized a smooth, AI-polished response to a classic market-sizing question, then froze completely when the interviewer added a follow-up constraint that wasn’t in his script. The candidates who did better were the ones who’d practiced the actual skill — breaking an unfamiliar problem into steps out loud, adjusting on the fly — rather than practicing a memorized answer to a specific question. That’s the real lesson: preparing with AI beforehand to understand structure is fine; trying to lean on it live is a fragile strategy that tends to collapse the moment a question deviates even slightly from the script.
How to Actually Prepare
Improving at brain teasers is less about memorizing answers and more about building a repeatable process: clarify the question, state assumptions out loud, break the problem into steps, and work through rough calculations transparently. Thinking aloud is the whole point — interviewers are evaluating the visible reasoning, not just the final number. Practicing with logic puzzles and reputable prep resources builds that muscle far more durably than rehearsing AI-generated scripts that fall apart under a live follow-up.
FAQs
Why are brain teasers still used in investment banking interviews if they don’t test finance knowledge?
Because they reveal how a candidate reasons under pressure and handles ambiguity — skills that matter more in real banking work than memorized formulas, and that are hard to fake with AI assistance in the moment.
Are banks aware candidates might use AI during interviews?
Yes — Goldman Sachs, Amazon, and other firms have issued explicit warnings against using generative AI tools during interviews, with disqualification as a stated consequence.
Can AI tools actually help answer a brain teaser in real time?
In theory yes, but the live, improvised, thinking-out-loud format makes it far harder to pass off an AI-generated answer convincingly than in a scripted technical question.
Is it cheating to prepare for brain teasers using AI beforehand?
No — using AI tools to understand structure and practice ahead of time is standard preparation; the line is using AI live, during the actual interview, without disclosure.
What skills do brain teasers actually measure?
Analytical thinking, clear communication of assumptions, creative or lateral problem-solving, and composure under uncertainty.
Should I expect more or fewer brain teasers in banking interviews going forward?
More, if the current trend holds — they’re one of the few interview formats that AI-cheating tools struggle to compromise, which makes them more useful to interviewers, not less.
Takeaway
Brain teasers were never really about ping-pong balls or gas stations. In an interview landscape where AI cheating tools can generate a textbook-perfect answer to a scripted question in seconds, the ability to reason visibly and adapt live is becoming one of the few things banks can still actually evaluate — which is exactly why brain teasers aren’t going anywhere.
Visit: Swifttech3
