Hallucination: a bitter pill to swallow
AI Today13 Touko 2025

Hallucination: a bitter pill to swallow

AI hallucinates 100% of the time. That's by design - without hallucinating the next word, this transformer architecture wouldn't exist.

Thankfully, LLMs built for general purpose applications are right 80% of the time. But that still leaves one in five outputs being questionable; not especially reassuring if you're an air traffic controller, or cardiologist.

How can we ever truly trust the machine?

On this episode of AI Today, we embark on a groundbreaking quest to ground these 'digital dreamers' in reality.

Discover how cutting-edge research is moving beyond just detecting the problem, to actively reducing the occurrence of incorrect hallucinations.

We delve into innovative techniques that employ internal fact-checking mechanisms, intelligently split complex queries to avoid confusing collisions, and meticulously track word-by-word groundedness against source material.

You'll learn how this confidence-boosting research is paving the way for the AI credibility revolution, a future where technology is not just remarkably powerful, but significantly more dependable.

Join us to understand the innovative solutions building AI you can rely on, where AI becomes trusted accelerator of success...

Jaksot(90)

Enzyme in plastic - it's FANTASTIC!

Enzyme in plastic - it's FANTASTIC!

Oh, Barbie Girl. Anyone remember Aqua? And in one fell swoop, the AI Today podcast lost 49 subscribers...Ready for a masterclass in competitive advantage?Picture this: AI’s next battleground isn’t jus...

19 Helmi 202518min

REVEALED: The truth about AI coding

REVEALED: The truth about AI coding

Imagine a world where software engineers are replaced by other software engineers that are entirely digital.No coffee breaks, no office politics, just pure, unadulterated code. It sounds like science ...

18 Helmi 202519min

Decoding Grok-3: Elon Musk's AI and the future of everything

Decoding Grok-3: Elon Musk's AI and the future of everything

This morning we woke to a new dawn for AI.Grok-3 is a cutting-edge large language model (LLM) from Elon Musk's XAI, designed to understand and generate human-like text, outperforming competitors in se...

18 Helmi 202514min

Table stakes: Businesses bet big on AI and data

Table stakes: Businesses bet big on AI and data

There are so many examples of ballsy businesses finding gems in their mountains of documents and datasets, thanks to the superpowered excavator that's AI, that cynics and sceptics are heading for the ...

17 Helmi 202523min

How to win with AI at work

How to win with AI at work

I've pulled together some really simple ways to make AI your best friend - or at least, frenemy - during the 9 to 5. This episode is dedicated to business leaders and everyone doing the real work. It'...

5 Helmi 202519min

The wildest month in genAI history...

The wildest month in genAI history...

What just happened? January 2025: a month of unprecedented change in generative AI. The open-source community is shaking up the industry, challenging the dominance of tech giants. This episode of AI T...

28 Tammi 202515min

42 years to build an app

42 years to build an app

This is the most personal episode I have ever recorded. It's my near half-century journey to building an app. And how AI made it possible. I'm a pre-beginner coder. Yet eventually - and I mean, eventu...

27 Tammi 202523min

25 extraordinary business AI use cases

25 extraordinary business AI use cases

Warning: After listening to this podcast, you won't be able to sit still. The possibilities revealed by AI will have you leaping into action, driving innovation, and revolutionising your business. Are...

16 Tammi 202530min