Asking ChatGPT Tough Privacy Questions

Подписаться 227 тыс.

Просмотров 9 тыс.

50% 1

How much does ChatGPT know about privacy & security? Let's find out!
🔐 Our Website: techlore.tech
🕵 Go Incognito Course - to learn about privacy: techlore.tech/goincognito
🏫 Techlore Coaching - to get direct support: techlore.tech/coaching
💻 Techlore Forum - to connect with other advocates: discuss.techlore.tech
🦣 Mastodon - to stay updated: social.lol/@techlore
We cannot provide our content without our Patrons, huge thanks to:
BRIGHTSIDE, Clark, Ente, Larry, Afonso, Boori, Brad, Casper, Cookie, Floyd, JohnnyO, kevin, love your content, NotSure, Poaclu, x
🧡 Join them on Patreon: / techlore
💚 To see our production gear, privacy tools we use, and other affiliates: techlore.tech/affiliates
💖 All Techlore Support Methods: techlore.tech/support
00:00 Intro
00:05 Is privacy important?
01:13 How to prevent data breaches?
02:17 VPN vs Tor?
03:20 Most private messengers?
04:32 Best password managers?
05:58 How to harden Firefox?
08:03 FOSS vs proprietary security?
09:59 How to escape government surveillance?
#chatgpt #privacy #techlore

Наука

Опубликовано:

29 июн 2024

Ссылка:

Скачать:

Готовим ссылку...

Добавить в:

Мой плейлист

Посмотреть позже

Комментарии : 41

@techlore Год назад

*Only 2 months late!* ❤If you enjoy our free content, check out our Patreon to help us grow: patreon.com/techlore

@RAM_845 Год назад

Bitwarden ftw

@AR15ORIGINAL Год назад

I also feel like ChatGPT forgot to mention something which tbf I'm one of the few who say: how citizen privacy and government transparency are two sides of the same coin, with govt transparency being necessary to guarantee democracy and citizen privacy protecting freedom.

@trappedcat3615 Год назад

1:02 Bing Chat excels here IMO. It provides citation, up to date info from the web, and does not require a phone number to use.

@Fr_g Год назад

I'm pretty sure the answers would be roughly the same, probably some more up to date info

@mr.mistoffelees7188 Год назад

Perhaps ask it about the "I have nothing to hide" statement and see how it fairs

@kueacybtguicyregfibubkueacybax Год назад

The reason you got some of the answers you got that you didn't like is because ChatGPT (the version you're using) is based on an older dataset, so for example, the lastpast recommendation is from an earlier time period where it was more common

@27370 Год назад

Really enjoyed this video! I wonder how much better GPT-4 would’ve handled these questions.

@ruiztulio Год назад

Consider that the tdata used for ChatGTP 4 is from year ago, some of the information is not up to date, I think that's why it recommends some out of date tools (like lastpass). Great video!

@MisterEightyFour Год назад

Is C-GPT only accurate up to September 2021? Would explain some of the omissions of key data?

@saisibi6708 Год назад

I want techlore to talk about privacy impacts from GPT models. Maybe show some Privacy friendly alternatives(if any).

@quickmythril2398 Год назад

i am curious if this was using the free version, or the GPT4 model available with the premium version. it can handle some more complex things in detail, like the multi-step process of hardening browser settings.

@oofsry Год назад

He’s using free version

@sinstreams Год назад

I’m gonna be honest, the windows antivirus is kinda the “worst” actual AV though it does protect against the more common/well known threats. Biggest issue is that most malware is pretty much programmed to disable the AV bc any file ran is in admin. Disabling the Microsoft AV is like common grounds for viruses so take that as you may. It’s a trade off if you want more security. Can’t have the best of both worlds all the time. Balance security and privacy where you don’t make too many sacrifices on either side. 1:48

@pueraeternus. Год назад

I'd ask it stuff like "is privacy unimportant" etc, I like to check everything for bias/telling me what I want to hear.

@timonelfert Год назад

Nice job using a current trend to get your message to a wider audience!

@pmjeterjr Год назад

Why is Enpass never mentioned for password managers. they do not store your passwords on their servers. They let you use storage of your Choice.

@redeyesdrogon786 Год назад

Great video!

@dbess1 Год назад

ChatGPT consistently confused the privacy w security. I wasted 45 mins looking for a list of 7 lightweight, security focused Linux distros. I asked for descriptions of each, download file size and display in order of file size least to most. The one I decided to try out, turned out to be discontinued in 2013! It listed Tails which isn’t security focused. It listed Qubes which isn’t lightweight. So for sure it’s not replacing other search engines and personal research. Not right now….hell apparently Openai is already being sued for slander for publishing false info via chatgpt. It’s definitely work in progress.

@8O8DRAX Год назад

nice review the quality drops yes sr.'

@lancemarchetti8673 Год назад

The fact that encryption means mere to nothing with your photos on your Google drive... I came up with my own privacy guard. No passwords etc, just Notepad++ Before upload, I simply insert/delete a Tab at certain points in the ASCII code of my image files, and save them. The image headers remain in tact but opening the image reveals a distorted unrecognizable image. A user would need to know where exactly the Tabs must be inserted or deleted to reveal the original image. Even online AI services have struggled to correct my jpeg code to reveal the original image.

@landotter Год назад

Good to know that human advice is still necessary. Great video.👍

@Hector-zj2ey 11 месяцев назад

What are some steps we can take to use ChatGPT more privately? For example, not sharing sensitive information and a VPN. Thank you

@I.____.....__...__ Год назад

- You should have asked if privacy is dead; it would be interesting to see what it would say. 🤔 - 2:33 ChatGPT said VPNs are used to access geo-fenced content but didn't mention that doing so is against the TOS of pretty much everything and grounds for suspension. It parroted VPN marketing materials. 🤦 - 3:27 WhatsApp is number 2? _Facebook's_ app? 🤨 - 3:29 Wickr can detect if you've taken a screenshot… only in Android. (Also, what good is it? It doesn't prevent or delete it.)

@maxim9733 Год назад

And also Telegram devs send some let's say "keys" (of course they will refuse that but the facts saying about it) to access to even private e2e chats for some countries governments, at least Russia with their FSB (like FBI in the USA). So: don't rely on Telegram. Especially if you live around of dictatorship countries. Signal is your choice but Signal require you must have a phone number existing, and that is potential flaw in your security. Session app is like Signal but without neediness having phone number at all.

@cameronmoore136 Год назад

This was fun

@chrisbowser Год назад

Problem for me is it answers everything like an American lawyer… well you could do X or you could do Y. I’m not going to advise you cos you might sue me… To me chatGPT and similar tools are like suped up Google searches

@peanut366 Год назад

Did you use your personal number to access ChatGPT? I'm very hesitant, and it won't allow me to use MySudo, Google Voice, or any other VoIP numbers.

@M0rn1n6St4r Год назад

1:43 *"6. Use a reputable antivirus program..."* - I am _pretty sure_ the *"built-in, native anti-malware support"* Henry mentions, _counts as_ *"reputable antivirus"* programs. It's not Henry's fault. Like the rest of us, he has become accustomed _to deciphering_ poorly worded, poorly spelled, poorly punctuated, run-on sentences with no main clause... generated by _natural intelligence._ Henry can't switch it off. Not even when ChatGPT writes _exactly what it means_ and means _exactly what it writes._ As opposed to _the norm_ (i.e. natural intelligence): "Don't read what I wrote. Read what _I meant to write."_ I wonder if that relates to the number and severity of _misinterpretations_ and _misunderstandings_ in our online interactions? Perhaps, could it be one of the causes?

@8O8DRAX Год назад

TECHLORE WHAT UP ' nice!!

@freeman788 Год назад

What about matrix over tor? And what happened on odysee? Why disable comments?

@kicker7955 Год назад

A language model is not Inteligent, not any more than a survey given to a crowd of people is. It just repeats what the most likely "corect" word is in the sentence. If you ask it: is facebook secure, it will give a completely different answer (probably referencing a paid audit and some PR articles or published papers) than if you ask it are most social media sites secure (the tone of which will reference probably hacker news and other security blogs). It doesn't decide, analyze or judge. It's just a search engine, a very advanced and very user engaging search engine, but ultimately, the brain, the inteligence behind the substance is just as dumb as Google 2008, before they "personalized" results. In fact, honestly, I think that's the biggest part of the appeal. If feels like how a search engine in the 21st century should look like. More gathering info, less delivering SEO bait and large publishers.

@4etri Год назад

I believe network security is something unreachable as of current state of artificial knowledge. I find it far too perplexed to even give you a final result. May be it can reach the point of giving you a graph stats and possible solutions. But would still require a knowledgeable engineer to analyze all of the info.

@danmar007 Год назад

It's amazing the number of people who think ChatGPT is great. I guess in a world of badly educated plebs, this kind of thing fascinates.

@likebot. Год назад

I'm going to stand firm on my belief that 2FA is not just unneccessary and arbitrarily removes your right to be responsible for yourself, but a serious liability if it uses an APP or any means of transmitting a pin for you to use. In other words, only a TOTP such as YubiKey could do. I'm also vehemently opposed to the practice of pairing a service to your device's on-line fingerprint (TeamViewer just implemented that grr).