Tech Talk: Scary, Scarier and just plain frightening
 | Author:
Vic Sample
Vic Sample: MT43 News Treasurer |
Tech Talk: Scary, Scarier and just plain frightening
Vic Sample
I have seen several articles lately that are just worrisome:
Zoom recently announced a new Terms Of Service with a clause that seems to grant them the right to your data (voice, video and messaging) to train their Artificial Intelligence model. After a firestorm of bad press, Zoom has released a revised version of the Terms of Service stating that the company does “not use customer audio, video or chat content without customer consent”. It’s hard to tell if the original version was just poorly written and gave the wrong impression or if Zoom changed its Terms Of Service due to the public backlash it received.
There are reports that Microsoft is planning to release a version of their Edge (the Microsoft web browser) that will take a screenshot of EVERY page you visit using Edge. The stated reason is so that you can revisit those sites even when you are offline. Microsoft did say they would not upload those screenshots to their databases. HMMM? Maybe – Maybe not.
And finally, researchers have found a way to train software to recognize what you are typing on your keyboard by the sounds the keys make. This spyware can exist on any device that has a microphone (your pc, notebook or phone) and listen to what you are typing.
While it is a frightening thought that any device in your house with a microphone could be listening and stealing your information, it’s unlikely to be a concern right now. The researchers used Zoom to record typing and trained it to recognize the individual keys on the pc keyboard being used. Not only would threat actors have to install the spyware on your device, they would have to train the spyware to recognize your keyboard keys.
It is one thing to do that training in a research lab; but a whole different thing to do it in the real world.
BUT, I remember when voice recognition was in its infancy. I visited a company that was working on developing voice recognition. The amount of “training” on a listening device to recognize one individual’s voice was daunting.
Now, voice recognition works straight out of the box without any training. Pam likes to use voice-to-text when texting from her phone. If I happen to be speaking it picks up both voices and merges everything together – generally into an unintelligible message – but it does do a good job of interpreting both voices a the same time.
Someday, someone may find a way to easily tell what you are typing without having to train the software. It’s not unimaginable!
Always, Always, Always beware out there!
