AI can give students answers that sound convincing but aren't always right.
That's why learning to question and verify AI-generated information is so important, especially when students are still learning the basics.
#FourierStudy#AILiteracy
@neyhaal96@justinskycak That's a good distinction. If a student can't remember something, it doesn't always mean they need more retrieval practice. Maybe they didn't fully understand it in the first place, and that's what needs addressing.
@ramtinrhmti Good that you're sharing the results either way. Just wondering, will you also test what happens when there's not enough information in the fault report? Would be interesting to see if the model says it doesn't know or just comes up with something that sounds right.
@ndgmlondon Good point. But what happens to anything that's already running or queued when the breaker kicks in? I imagine stopping new activity is one thing, but also making sure everything else has stopped could be another challenge.
@algoretic You raise good points here. I think knowing when to step in is the difficult part. Some students might need another hint, while others need the problem broken down a bit more. Getting that balance right seems just as important as knowing when to hold back the answer.
@ComplexD That's quite concerning, especially when companies are uploading confidential information. You'd want to know who's actually handling that data and whether the promise not to retain it applies to everyone involved, not just the app you're using.
@iam_tsr I wonder what happens when a number has a footnote or a unit somewhere else in the table. The AI might pick up the right number but still get the answer wrong because it's missed what that number actually means.
@BlueTutorsUK And when you get stuck halfway through explaining something, you know exactly what you need to go back over. Worth trying it again a few days later too, just to see if it's actually sunk in.
A supplier you checked last year might look very different today.
Businesses change, supply chains shift and new risks can emerge.
That's why due diligence shouldn't be a one-off exercise. It's something worth revisiting regularly.
#CSDDD
It's easy to focus on the big clauses when reviewing a contract. But sometimes it's the smaller details that end up costing you later.
Before you sign, make sure you know what you're agreeing to.
Try ContractIQ free:
https://t.co/NQPlMqQ0H9
@SiuHarry202601 And even when someone approves an action, the AI should still be limited to exactly what was approved. If the target or environment changes, surely it needs to ask again?
@kritikandhari I think that's the challenge. If students are used to getting answers straight away, being asked to work through something might feel frustrating at first. I suppose a lot depends on how well the AI guides them without making it feel like extra homework.
@RSABIVARMA That's the bit people might miss. The AI hasn't made up the number, but it's still giving the wrong answer. And unless someone checks where that number came from and what it actually means, it could easily go unnoticed.
@BigPapaCluckk@SciInTheMaking Clever! At least the students are having to think about what they've just read or watched, rather than copying it all down. Do they find it difficult at first, remembering things without looking back?
@XpatEducator I hadn't thought about silent recall being more useful for simpler things. I suppose it comes down to knowing when students need to write something down or explain it out loud, rather than just thinking they've remembered it.
Revenue is growing. Great.
But are your margins growing with it?
Discounts, pricing and cost-to-serve can quietly eat into profitability.
Find out where to look with our free Margin Scorecard.
Comment MARGIN and we'll send it over.
@tshokama And sometimes it finds the right document but misses something important. Like picking up the main rule but missing an exception further down. The answer sounds right, but that one missing detail could change everything.
@dheeraja486 I suppose the risk is that an agent remembers what it still needs to do, but loses track of why something was left unfinished in the first place. And that could be a real problem when dealing with legal documents.
@CKPillaiAI Right, and even when the source is reliable, it's worth checking whether it actually backs up what the AI is saying. Sometimes the source is perfectly fine, but the AI has drawn the wrong conclusion from it.