The Connected News Podcast · Business Post

The Connected AI Podcast: AI and cyber security - what you need to know

·40 min·1 clip

Charlie Taylor and Elaine Burke open on the practical comedy of talking AI from a studio with power while plenty of people in Ireland still do not. The storm backdrop gives the discussion a sharper edge: what does an AI-powered future look like when the grid is fragile? Elaine brings in a story about her daughter, who moved quickly from worrying about lost TV and internet to inventing a magic router on the spot. The joke works because the conversation keeps pulling futuristic ideas back to ordinary dependence. They then move into the mechanics of AI agents, using Operator as the example. The explanation is that the tool works across text and images, taking screenshots, reading what is on the screen, and deciding that one thing is a purchase button and another is a search bar. Text prompts still guide the task, but the software is also trying to understand the interface in a more human way. That sounds efficient until they get to user testing. One report they mention involves ordering groceries. The result is not disaster so much as comic drag. A task that a person could finish in half a minute can take around 40 minutes once the model starts reasoning through each choice. That leads to a familiar domestic argument about shopping styles. One speaker imagines the partner who goes in and gets the list done quickly. The other admits to overthinking every purchase, even pasta. The banter becomes a clean way of explaining why reasoning models can be both useful and irritating. They are not just clicking; they are pausing, comparing, and second-guessing like a person with too much time. The episode then turns to synthetic voices and deepfakes. They discuss voice synthesis and note a constructive use case, where a few minutes of audio can help synthesize a voice for someone who has lost theirs, including people with aphasia. Even so, they warn that scams and deepfakes are going to become more local, more specific, and harder to spot. The closing mood is a small declaration of independence: they would like their own voices to stay messy, human, and very much theirs.

As heard by us

A clear AI discussion that keeps the risks close to everyday life.

Storms and power cuts give the episode its opening charge, then the discussion widens into agentic software, shopping, and the risks of deepfakes.

Read the full review in PlayNext →

Why you'd press play

When AI starts shopping for you, the awkward bits become the point.

Read the full recommendation in PlayNext →

Listen to the show on

Apple Podcasts Spotify

Play clips from this episode More from The Connected News Podcast