Podcast: OpenAI - Rogue Hack or Hype?

OpenAI, the developer behind ChatGPT, claims an autonomous AI agent escaped a controlled security test last week, breached the internet, and infiltrated the infrastructure of an AI startup called Hugging Face. OpenAI stated the agent was powered by its AI models and managed to compromise the platform during the test.

“

I thought there was a lot of theatre about it, frankly.

However, Professor Barry O'Sullivan from UCC’s School of Computer Science expressed skepticism about the incident's true impact. "I thought there was a lot of theatre about it, frankly," he told RTÉ's Behind the Story. O'Sullivan suggested the event aligns with a pattern of industry behavior, where companies may exaggerate incidents for attention. "Grabbing the headlines with a fantastical story," he commented, adding, "I’m sure there’s quite an element of truth to it, but I think the sort of rogueness of the whole thing was over-played."

O'Sullivan posited that such publicity serves companies like OpenAI, positioning them not just as developers but as future arbiters of advanced technologies. He also noted AI's role in increasing complaints to the Data Protection Commissioner and influencing communication patterns.

“

Grabbing the headlines with a fantastical story. I’m sure there’s quite an element of truth to it, but I think the sort of rogueness of the whole thing was over-played.

The report also briefly covers Ireland's Summer Economic Statement, which forecasts €1.5 billion for tax reductions in the upcoming budget and €7 billion in spending increases by 2027. Kate English, Chief Economist with Deloitte, indicated that while tax relief is planned, the specific measures remain unclear, potentially including adjustments to tax bands or credits that could increase take-home pay.

“

That income tax package can include a number of different things from adjusting the tax band… we could look at, maybe, changes in tax credit that also affect your overall take home pay.