Conversation

Notices

1. Embed this notice
   Rich Felker (dalias@hachyderm.io)'s status on Friday, 18-Apr-2025 23:05:51 JST Rich Felker

   • Kevin Beaumont

   @GossiTheDog Hey I'm just telling them what they should do if they want to be competitive or even still viable once this shit melts down.

   Their loss if they don't want to listen to me.

   See also: my display name.

   • Embed this notice
     Kevin Beaumont (gossithedog@cyberplace.social)'s status on Friday, 18-Apr-2025 19:03:01 JST Kevin Beaumont

     LinkedIn is sadly a really good sense check for where businesses are going.

     One thing I’ve noticed lately - people come up with the stupidest ideas possible, and then post long things about how it’s a great idea. They use generative AI to make the posts - generative AI will write posts and documents supporting whatever dumb shit you want.

     Rich Felker repeated this.
   • Embed this notice
     Kevin Beaumont (gossithedog@cyberplace.social)'s status on Friday, 18-Apr-2025 19:04:32 JST Kevin Beaumont

     in reply to

     My feeling - this will transfer inside businesses. People will present long Word documents and PowerPoints proposing absolutely dumb shit, and because everybody in biz is so numb to the process, they’ll nod along to the green tick lists and 57 page Word documents and sign things off which shouldn’t be.

     Rich Felker repeated this.
   • Embed this notice
     Kevin Beaumont (gossithedog@cyberplace.social)'s status on Friday, 18-Apr-2025 19:10:45 JST Kevin Beaumont

     in reply to

     Automating people being able to express dumb ideas at scale with long documents full of errors and made up data will have consequences, basically - as people don’t have the bandwidth to refute infinite bullshit.

     Blurry Moon likes this.
   • Embed this notice
     Kevin Beaumont (gossithedog@cyberplace.social)'s status on Friday, 18-Apr-2025 19:16:28 JST Kevin Beaumont

     in reply to

     IMHO businesses should have a global policy for every employee saying all generative AI content in documents - Word, PowerPoint slides etc - must be clearly labelled as such.

     That way when you’re reviewing documents to make decisions, you have an idea around the source.

     LukeAlmighty 🇨🇿 repeated this.
   • Embed this notice
     Kevin Beaumont (gossithedog@cyberplace.social)'s status on Friday, 18-Apr-2025 19:35:30 JST Kevin Beaumont

     in reply to

     As an example - did you know that you don’t need a functioning CVE system as Copilot says you can buy magic AI boxes? Everything is automated, job solved. People in cyber are posting this kind of dribble on LinkedIn.

     Phantasm likes this.
   • Embed this notice
     mr_harm (mr_harm@mastodon.social)'s status on Friday, 18-Apr-2025 19:57:10 JST mr_harm

     in reply to

     • Kevin Beaumont

     @GossiTheDog I’m not a cybersecurity specialist by any means and I tend to ignore LinkedIn because even before AI it was an unbearable place of self-presentation and bragging about minor things.
I see AI as “base”-protection: if used right, even a beginner in cybersecurity can achieve a good protection against well-known threats by prompting and implementing with a bit of common sense. So basically it would free professionals from the basics to focus on more important tasks

   • Embed this notice
     Blurry Moon (sun@shitposter.world)'s status on Friday, 18-Apr-2025 20:49:09 JST Blurry Moon

     in reply to

     • Kevin Beaumont
     • Simon Zerafa
     @simonzerafa @GossiTheDog with AI it takes no effort to create bollocks so the formula has to be updated
   • Embed this notice
     Simon Zerafa (simonzerafa@infosec.exchange)'s status on Friday, 18-Apr-2025 20:49:10 JST Simon Zerafa

     in reply to

     • Kevin Beaumont

     @GossiTheDog

     It's takes twice as much energy to refute bollocks than it does to create it.

     Therefore no-one will want to take the time, energy and other resources to do so.

   • Embed this notice
     Jukka Niiranen (jukkan@mstdn.social)'s status on Friday, 18-Apr-2025 20:51:46 JST Jukka Niiranen

     in reply to

     • Kevin Beaumont

     @GossiTheDog Once a human prompts Copilot to create a doc and then saves it to M365, that now becomes a human authored source. Next time that hallucinated BS will be attributed to a coworker, making it even harder for anyone to spot the error.

     I've asked if there is anything built into #Microsoft365 #Copilot to stop this feedback loop of AI hallucinations. No, there isn't.🤷♂️

     LukeAlmighty 🇨🇿 and Phantasm like this.
     LukeAlmighty 🇨🇿 repeated this.
   • Embed this notice
     Encyclopædia Autistica (Order Now!) (mrsaturday@shitposter.world)'s status on Friday, 18-Apr-2025 20:52:04 JST Encyclopædia Autistica (Order Now!)

     in reply to

     • Jukka Niiranen
     If you ever get bored, have Microsoft 365 access, and want to have a little fun at Microsoft's expense, put blatant misinformation or opinions meant to nudge Clippy 2.0 into shilling on your behalf on their servers
     LukeAlmighty 🇨🇿 likes this.
   • Embed this notice
     LukeAlmighty 🇨🇿 (lukealmighty@gameliberty.club)'s status on Friday, 18-Apr-2025 20:53:48 JST LukeAlmighty 🇨🇿

     in reply to

     • Encyclopædia Autistica (Order Now!)
     • Jukka Niiranen

     @mrsaturday @jukkan
     Oh, you think Master/ Slave was bad?

     Just wait till Copilot starts using Ubermench/filthy_degenerate functions...

   • Embed this notice
     Phantasm (phnt@fluffytail.org)'s status on Friday, 18-Apr-2025 20:57:37 JST Phantasm

     in reply to

     • Kevin Beaumont
     @GossiTheDog To be fair, I see this as a good thing. Businesses run by people that are more than happy to sign off on stupid and dumb things shouldn't be businesses for long. They'll get run into the ground by their own stupidity, lack of perception grounded in reality and get filtered by businesses run by people that aren't afraid to think in the corporate environment.
   • Embed this notice
     Rich Felker (dalias@hachyderm.io)'s status on Friday, 18-Apr-2025 23:03:50 JST Rich Felker

     in reply to

     • Kevin Beaumont

     @GossiTheDog Businesses should have a global policy for every employee that generative AI content in documents...

     ...is grounds for immediate dismissal.

   • Embed this notice
     ⠠⠵ avuko (avuko@infosec.exchange)'s status on Friday, 18-Apr-2025 23:04:11 JST ⠠⠵ avuko

     in reply to

     • Kevin Beaumont

     @GossiTheDog from experience: this is already happening.

     Not only subtly bullshitty stuff about strategy etc, but I’ve even seen a case where technical documentation about features and APIs was generated by chatGPT and copy-pasted.

     So with zero factual content, and zero attribution or labelling as “pure fantasy”. Took the team weeks to recover.

     People worry about Confidentiality, but the biggest threat of GenAI is to Integrity.

     PS: Yes, I like the summary functionality (although I always vet carefully and end up rephrasing most), and as a non-native speaker starting with something generated really helps sometimes.

     But for the love of all of our collective history, stop the disinformation pollution.

   • Embed this notice
     Rich Felker (dalias@hachyderm.io)'s status on Friday, 18-Apr-2025 23:04:54 JST Rich Felker

     in reply to

     • ⠠⠵ avuko
     • Kevin Beaumont

     @avuko @GossiTheDog "People worry about Confidentiality, but the biggest threat of GenAI is to Integrity."

     👆 🛎️ 🛎️ 🛎️

   • Embed this notice
     Julian Wolf (julupu@infosec.exchange)'s status on Saturday, 19-Apr-2025 02:53:20 JST Julian Wolf

     in reply to

     • Kevin Beaumont

     @GossiTheDog I've seen kind of the opposite policy at my last employer (fortune 50 company). Every department, every employee needs to use GenAI - the more the better. It's all bullshit top to bottom.

   • Embed this notice
     Death by Lambda (xdydx@mastodon.social)'s status on Saturday, 19-Apr-2025 06:35:37 JST Death by Lambda

     in reply to

     • Kevin Beaumont

     @dalias @GossiTheDog
     Cass is so write. I'm predicting my day-rate will more than double in 2 years as companies need #AIExorcism.

     I would love to know, just on LinkedIn, how many inappropriate interviews have taken place because of the recruiter and the candidate having only corresponded through AI..