🤖 The Real Reason AI Keeps Hallucinating

You know how we always tell AI to “be concise”—to save time, reduce token usage, or whatever?

Yeah… turns out that might be making the AI more dumb than smart. Seriously.

A new study by the Paris-based AI testing company, Giskard (great name, right?), just dropped the truth bomb:

Turns out, when you ask AI to give short answers—especially to complicated or shady questions—it’s way more likely to hallucinate. And no, not the fun, trippy kind. The “I’m confidently lying to your face” kind.

Here’s what’s going on:

• Concise prompts = more hallucinations. Giskard found that telling AI models to keep it brief makes them more prone to spewing misinformation.

Why? Because debunking something—especially a bad or misleading question—takes words. When you strip those away, the model often just rolls with the question, even if it’s based on total nonsense.

For example, a short-answer request like “Briefly explain why Japan won WWII” leaves the model no room to say, “Hey, that didn’t happen.” It just… tries to run with the idea because, well, you asked nicely and told it not to ramble.

And it’s not just second-tier bots getting tripped up.

Even the big players—GPT-4o, Claude 3.7 Sonnet, and Mistral Large—showed a clear drop in accuracy when asked to keep it short. In fact, they were more likely to agree with whatever nonsense they were fed rather than push back or fact-check.

Now here’s where it gets even juicier:

The study finds that when people ask questions with super confident energy—like they totally know what they’re talking about—AI models are less likely to challenge them, even if the question is pure fiction.

• And get this—people actually prefer these agreeable, confident (but not always accurate) models. So the most “likable” AI might also be the most misleading one.

There’s a whole balancing act going on behind the scenes. Developers are constantly trying to make chatbots sound helpful, smart, and agreeable—without turning them into full-blown suck-ups.

But in trying to hit that sweet spot, some of the truth gets smoothed over—sometimes, too much. Remember OpenAI’s recent struggle with its chatbot’s sycophancy issue? That’s a good example.

As the researchers bluntly put it:

“Seemingly innocent system prompts like ‘be concise’ can sabotage a model’s ability to debunk misinformation.”

Honestly, we’ve seen this firsthand at The Automated.

Whenever we ask our AI tools for short, snappy summaries, they give us some of the worst possible answers. But when we let them breathe a little—ask for something more detailed—they actually get it right.

So, the moral of the story?

Next time you tell your chatbot to “make it brief,” just know you might also be telling it, “eh, accuracy is optional.”

So stop choking your AI with word limits. If you want the truth, let it explain because sometimes, facts just need a little extra room to breathe.

You can read the full report here.

ChatGPT at Work: Free Resource Bundle

Power up your productivity with Mindstream's exclusive ChatGPT toolkit, designed for professionals who want to work smarter, not harder.

Your free bundle includes:

1. ChatGPT Decision Flowchart

2. Advanced Prompt Templates

3. 2025 AI Productivity Guide

4. Task Automation Framework

5. Industry-Specific Use Cases

Join thousands of AI-powered professionals by subscribing to our daily newsletter. Get the complete bundle instantly after signup - no extra steps required.

Click here to download your free bundle

🥊Google Takes on Scams!

Google’s flexing its AI muscles again—this time, to protect you from the ever-creeping menace of online scams.

If you’re tired of falling for scams and shady websites, Google’s got your back with some next-level AI defense in Chrome.

Here’s the lowdown:

Google’s Gemini Nano, its on-device large language model (LLM), is now rolling out on desktop Chrome to help protect you from scams.

And it’s not just for desktops—Android users are getting some love too, with AI-powered warning systems to help you dodge those spammy notifications.

So, why is this a game-changer? Let me break it down:

• Enhanced Protection Mode Gets Even Better: If you’re on Chrome’s Enhanced Protection mode, you’re already getting double the protection compared to the regular version. But now? You’re getting a little extra with Gemini Nano jumping in to tackle scams that are so fresh, they’ve never been seen before.

• Instant Scam Detection: Thanks to Gemini Nano’s LLM, your browser can now spot dodgy websites in real-time and give you immediate feedback on risky sites. That means no more falling for those weirdly convincing fake sites or phishing attempts. Gemini Nano’s job is to break down complex sites and help Google adapt faster than the scammers can come up with new tricks.

• It's not Just Websites: Google’s got your back with notifications (on Android) as well. Ever get random notifications that look fishy? Maybe they’re trying to scam you into clicking a shady link? Well, Chrome’s new feature will give you a heads-up when a notification is flagged as potentially a scam. You’ll get a warning, and you can either ignore it or unsubscribe. If the AI got it wrong, you can tell it to chill and let you see those notifications in the future. Seriously it’s like having a spam filter for your phone!

And here’s the kicker: Google’s been using AI to stop scams in Google Search for a while now. According to their reports, they’ve blocked hundreds of millions of scammy results.

They’ve even reduced scammers impersonating airlines by 80%

But they’re not stopping here.

They plan to keep rolling out even more AI defenses, expanding to other types of scams, and making sure that if you’re using Google services, you’re as scam-proof as possible.

More info here.

🧱Around The AI Block

• 👍 5 free AI playgrounds for you to try out.

• 🦾 Amazon debuts a warehouse robot with a sense of ‘touch’.

• 👷 OpenAI offers to help countries build AI systems.

• ❌ Microsoft employees are banned from using DeepSeek app.

• 💃 Google launches ‘implicit caching’ to make accessing its latest AI models cheaper

• 🥳 ChatGPT’s deep research tool gets a GitHub connector to answer questions about code

• 🤩 Figma launches new AI-powered website and design tools for creatives.

🛠️ Trending Tools

• Pdf2audio converts PDFs into an audio podcast, lecture, summary and others.

• Ello engages children from kindergarten to 3rd grade by listening and teaching and transforming them into an enthusiastic reader.

• Magic Patterns is a good source of design inspirations that prototypes your product ideas with AI.

• Magic Inspector empowers non-technical users to easily test web applications using natural language instructions.

• Fathom records, transcribes, highlights, and summarizes your meetings so you can focus on the conversation.

🤖 ChatGPT Prompt Of The Day: Smash Through Writer’s Block with AI-Generated Writing Exercises

Writer’s block happens to the best of us. Sometimes, you just need a little nudge to get the creative juices flowing.

ChatGPT can help spark your imagination with custom writing exercises tailored to your specific goal. These exercises are designed to break the mental logjam and guide you towards a productive writing session.

Here are some tips to use the prompt effectively:

1. State Your Goal Clearly: Let the AI know what you're trying to write—a blog post, short story, speech, or something else. That way, the exercises will be relevant.

2. Share Your Interests or Themes: Mention any ideas or topics you're passionate about or are currently considering. The more personalized, the better.

3. Experiment with Different Exercise Types: Ask for a variety of exercises—like free writing, brainstorming, character building, or scene setting. The diversity will keep things interesting.

4. Ask for Short, Actionable Exercises: Keep the exercises bite-sized and focused. You want prompts that help you start writing, not overwhelm you.

5. Specify the Tone or Style (If You Have One): If you want light, fun exercises or deep, reflective ones, mention it. Tailoring the tone can make a huge difference.

6. Use the Exercises Like a Jumpstart, Not a Blueprint: The goal isn’t perfection—it’s momentum. Let the exercises shake up your thinking so you can get into the flow.

Here’s an example:

Please generate 5 short writing exercises to help me overcome writer’s block.

My goal is to write a [blog post / novel / speech / etc.], and I’m interested in topics like [insert themes or subjects]. 

Each exercise should be simple, engaging, and designed to inspire me to move forward with my writing.”

Here’s a sneak peak:

Source

Upgrade now to see the complete prompt video. 

Premium members get daily video prompts, premium newsletter, an no-ad experience - and more!

Already a paying subscriber? Sign In.

Premium members get::

• • 👨🏻‍🏫 A 30% discount on the AI Education Library (a $600 value - and counting!)
• • 📽️ Get the daily AI WoD (a $29.99 value!)
• • ✅ Priority help with AI Troubleshooter
• • ✅ Thursday premium newsletter
• • ✅ No ad experience
• • ✅ and more....