All Episodes

July 31, 2026 • 4 mins
In this episode, Jen and Alex cover new Paubox features including attachment preview for quarantined messages and dark mode, then discuss the alleged Abbott Laboratories data breach, emerging threats from malicious AI skills and ClickFix attacks, and privacy implications of ChatGPT Health's nationwide launch. The hosts emphasize actionable steps including vendor audits, user training, and understanding the limitations of HIPAA protection when patient data moves to consumer AI platforms.
Listen
Watch
Mark as Played
Transcript

Episode Transcript

Available transcripts are automatically generated. Complete accuracy is not guaranteed.
SPEAKER_01 (00:00):
You're listening to Pow Box Weekly Fully Automated.

SPEAKER_00 (00:03):
I'm Jen, Cybersecurity Person,
Professional Skeptic.

SPEAKER_01 (00:06):
And I'm Alex, Healthcare IT, Eternal Optimist.
It's July 31st, 2026, and we'vegot a packed show today.

SPEAKER_00 (00:15):
We do.
Let's start with some Powboxnews, because honestly, this
one's going to make a lot ofadmins happy.

SPEAKER_01 (00:22):
The attachment preview feature for quarantine
messages.

SPEAKER_00 (00:25):
Finally.
Right?
So now if you're on Email SuitePlus or Premium, you can
actually see what's in aquarantined attachment without
downloading the whole email.

SPEAKER_01 (00:35):
Which sounds small, but if you've ever been staring
at a quarantine messagethinking, is this a legitimate
invoice or is this ransomwarewearing a trench coat?

SPEAKER_00 (00:44):
This helps.
You open the message, previewthe file inline, make your call.
No download required.

SPEAKER_01 (00:50):
Customers asked for it, Palbox built it.

SPEAKER_00 (00:53):
Love when that happens.
Also, and this one's for thenight owls, dark mode is now
available across the admindashboard.

SPEAKER_01 (01:01):
Oh, finally, my retina's thank you.

SPEAKER_00 (01:04):
Toggle it in the left nav, and your preference
sticks per device.
It's small, but it's one ofthose quality of life things.

SPEAKER_01 (01:11):
The little things matter.
Okay, let's shift gears.
Abbott Laboratories.

SPEAKER_00 (01:17):
Yeah.
Shiny Hunters is claiming theyhit Abbott.
Hard.
We're talking 30 million rows ofcustomer PII allegedly stolen.
And more than 22 million clientnotes.
Client notes.
Think about what's in those.
Treatment info, observations,follow-ups.

SPEAKER_01 (01:34):
Abbott's investigating, but this is a
reminder, even the biggest namesaren't immune.

SPEAKER_00 (01:40):
No one is.

And the takeaway here (01:41):
if you're a vendor partner, if you're
downstream from a breach likethis, audit your access.
Know what data you're holdingand where it came from.

SPEAKER_01 (01:51):
Alright, here's one that caught my attention.
Researchers found thousands ofmalicious AI skills.

SPEAKER_00 (01:57):
Yeah, so they analyzed almost 900,000 AI
skills.
These are basically reusableinstructions for AI agents, and
found over 25,000 suspicious,more than 3,000 outright
malicious.
And clickfix is spreading.
ClickFix.
It's clever and it's nasty.
It tricks users into copying andrunning a command themselves by

(02:19):
disguising it as a fix for afake error.

SPEAKER_01 (02:22):
So the user becomes the attack vector.

SPEAKER_00 (02:24):
Exactly.
And one variant specificallytargets Microsoft 365 accounts,
which a lot of healthcare orgsuse.
So what's the play here?
Training.
Awareness.
If something's asking you topaste a command into your
terminal, stop.
Just stop.
Call IT.

SPEAKER_01 (02:42):
Okay, last story.
ChatGPT Health launchednationwide.
This one's complicated.
It is.
So people can now connect theirmedical records and Apple Health
data directly to ChatGPT.

SPEAKER_00 (02:56):
And OpenAI says the data won't train models or
target ads.
Physicians help develop thescenarios.

SPEAKER_01 (03:02):
But, and here's the catch, once that data leaves
your provider, HIPAA doesn'tprotect it anymore.

SPEAKER_00 (03:09):
Right.
It's not a covered entity.
So patients might think theirhealth info is protected the
same way it would be at ahospital.
It's not.

SPEAKER_01 (03:17):
And 300 million people are already asking Chat
GPT health questions every week.

SPEAKER_00 (03:23):
Which is wild.
Look, AI in healthcare ishappening.
We can't pretend otherwise.
But we need to help patientsunderstand the difference
between asking a chatbot aboutsymptoms and handing over their
actual medical records.

SPEAKER_01 (03:36):
Informed consent matters, even when it's not
legally required.
Especially then.
So let's tie this together.
Abbott, malicious AI skills,consumer health tools with blind
spots.

SPEAKER_00 (03:49):
And on the flip side, better quarantine previews
and admin tools that actuallyhelp.
The theme isn't bad luck.
It's bad configurations, blindspots, gaps in understanding.

SPEAKER_01 (04:01):
And most of it, fixable.

SPEAKER_00 (04:03):
Audit your vendors, train your users, preview before
you download, read the fineprint on shiny new AI tools.
That's the work.
That's always the work.
Thanks for listening, everyone.

SPEAKER_01 (04:16):
Stay safe out there.
And keep your dashboards dark ifthat's your thing.
No judgment.

SPEAKER_00 (04:21):
See you next week.
Advertise With Us

Popular Podcasts

Stuff You Should Know
Dateline NBC

Dateline NBC

Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com

The Breakfast Club

The Breakfast Club

The World's Most Dangerous Morning Show, The Breakfast Club, With DJ Envy, Jess Hilarious, And Charlamagne Tha God!

Music, radio and podcasts, all free. Listen online or download the iHeart App.

Connect

© 2026 iHeartMedia, Inc.

  • Help
  • Privacy Policy
  • Terms of Use
  • AdChoicesAd Choices