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)

World-class customer research - without the customers!

World-class customer research - without the customers!

If you're familiar with the Eisenhower matrix you'll be familiar with businesses and customer research - they simply don't know what they don't know! But thanks to two crucial AI research studies, we'...

16 Tammi 202512min

2025: AI and what I'm building

2025: AI and what I'm building

The past few weeks in AI have shattered my brain into a billion fragments of wonder. We've even found a new way to do AI, beyond transformers - that could change even what's been the most changeful we...

23 Joulu 202410min

BIG LAUNCHES: Devin and Gemini 2 Flash overshadow Santa's sack

BIG LAUNCHES: Devin and Gemini 2 Flash overshadow Santa's sack

You've heard the pandemonium all about Google launching its fastest and smartest frontier model yet. But what does it mean for your business? And what about Devin - the grown-up AI copilot for your en...

12 Joulu 202418min

AI Tomorrow

AI Tomorrow

Finally - an overdue appearance from AI Today creator, Dave Thackeray! What a year it's been. And it's just the beginning. Join me taking a look at 2024 and the indisputable delights and miracles co...

9 Joulu 202414min

EXCLUSIVE: AI gets memory!

EXCLUSIVE: AI gets memory!

The last barrier to enterprise adoption of AI was memory. Baking into every prompt what the algorithm needed to know, was enough to send business leaders scurrying for the Luddite hills. But now Googl...

20 Marras 202427min

CMO wet dream: Predicting human behaviour

CMO wet dream: Predicting human behaviour

Understanding human behaviour is critical to business success. Behavioural science informs every growth stage and product decision - yet so few businesses pay any attention to human behaviour and psyc...

7 Marras 202416min

Accelerating R&D with AI

Accelerating R&D with AI

Tired of research and development (R&D) bottlenecks? Today's episode of AI Today explores how AI can supercharge product development by rapidly uncovering game-changing insights from mountains of data...

1 Marras 202422min

AI as your full-stack engineer: with Databutton, it's finally time!

AI as your full-stack engineer: with Databutton, it's finally time!

I've tested 20 AI coding editors. My tech skills are basic, at best. None turned my ideas into apps. That's when I found Databutton. And now I'm an app developer. Listen in to find out how Databutton ...

30 Loka 202427min