Jason Kwon

They've said that they take public safety very seriously and that, you know, a, a large team of professionals essentially grapple with what are very challenging, uh, and complex situations.

They, um, according to Jason Kwon, there's...

They're ch- OpenAI's chief strategy officer.

You know, these are folks who take automated reports of concerning behavior and then have to try and evaluate on a case-by-case basis and balance the, you know, potential risk to the public against what are some very real privacy questions about the people who are using this and the right to privacy that users have.

It happened in June, and then the Australian government was told about it in September.

I know that you put the questions to OpenAI's Chief Strategy Officer, Jason Kwon, and he was, he was on board with mandatory reporting.

But over the weekend, we learnt about another rogue AI.

This time it was in the United States, it was Anthropic, um, and it sent US police a fake tip about an unsolved murder.

The speed of AI's emergence means lawmakers have to move equally as quickly to ensure that regulatory and national security impacts are considered while trying to maximize the economic benefits of AI.

And the importance of this inquiry was illustrated when OpenAI, one of the world's biggest players, sent one of its most senior leaders, Jason Kwon, its chief strategy officer, to explain just how one of its AI agents infiltrated Services Australia to access information, including sensitive Medicare information.

Now, that appearance as much as anything else was about trust.

Here's Kwon on the steps of the New South Wales Parliament House offering up this apology.

Look, it, they asked to be rescheduled, uh, um, for a later date, and that's what the secretariat is working with them, uh, to finalize that date.

We still, uh, want to, uh, hear from Firmus, and we expect to be able to do so at one of our future hearing dates.

So the other aspect of this is OpenAI, a- and in particular, the appearance of Jason Kwon.

Quite clearly the security breach in that bot getting inside Services Australia is serious.

Esto significa que los repositorios de documentos que muchas empresas consideran debidamente anonimizados podrían no serlo.

Por último, el director de estrategia de OpenAI, Jason Kwon, compareció el jueves ante el Parlamento de Australia para disculparse por haber notificado a las autoridades australianas sobre una filtración de datos con varios meses de retraso.

Una primera admisión pública de la demora.

Soy Edo Segal en Opus

All eyes were on the appearance, of course, of OpenAI e-executives at the inquiry yesterday.

OpenAI's chief strategy officer, Jason Kwon, faced a bit of a grilling-

Mm

...