Full transcript
0:00Hello everyone. I don't normally make
0:02videos like this, but I wanted to since
0:06I have at least a small platform in this
0:09AI world, I want to make sure I give
0:12voice to something that is very
0:14concerning to me as a creator and
0:17someone who uses these AI tools every
0:19day. Coming from what is unfortunately
0:22perplexity. Now, if you followed this
0:25channel for a long time, you'll notice
0:27that I love perplexity. I has been my
0:30go-to tool for all of my like nonwriting
0:34tasks. I use it in place of Google
0:36search. I use it for just random stuff.
0:38I use it for troubleshooting stuff. But
0:41recently, they made some changes to
0:43their terms of services that almost went
0:45under the radar because they didn't
0:46announce it to people. That's the first
0:48red flag right there is that they didn't
0:51announce it. Uh at least from what I'm
0:53told, they didn't announce it. And there
0:55are a number of very concerning things
0:58in those terms of services that make it
1:00basically unusable for me if they don't
1:02change it. So, I'm making this video as
1:04a way to call attention to these changes
1:06and hopefully to inspire more people to
1:10contact Perplexity and let them know
1:13that you don't like these things cuz I
1:15would be really ashamed for any of these
1:17things to continue in Perplexity. They
1:19would cause me to switch to a different
1:21provider. I don't want that to happen
1:22because I love Perplexity. It's it has
1:25been up to this point my favorite
1:27chatbot overall compared to the chat
1:29bots of Chat GBT or Claude. It has
1:32things like being able to use multiple
1:33models within the same chatbot which you
1:36can't do with any of the other major
1:37ones. So, let's actually dive in and see
1:40what these new terms of services are.
1:42First of all, you can access them
1:45directly from Perplexity's legal hub.
1:47I'll make sure to leave a link to this
1:49in the description of this video. These
1:52were last updated on January 23rd and
1:55unfortunately it's taken this this long
1:57to figure out what happened. You know, I
2:00am not a lawyer. Let's just put that
2:01caveat out there right now. I, you know,
2:04consult a lawyer if you want anything
2:06certain on any of these topics. But I
2:08would also like to thank this person
2:10called okay smell453
2:13on Reddit who went through and basically
2:15summarized all of the major changes from
2:18this January 23rd 2026 uh update. And
2:22there are a number of things in here. I
2:24just want to thank this person because
2:25he already did a lot of the hard work of
2:27actually extracting all of the b the
2:29main red flags. I did do enough research
2:32to make sure that the biggest of the red
2:34flags that I saw, I went and confirmed
2:36for myself that they are indeed in the
2:39terms of services. So, let's look at
2:41what some of these uh say. For instance,
2:46first of all, this is personal
2:48non-commercial use only. Despite pro
2:50branding suggesting professional and
2:52business capabilities, the terms
2:53explicitly restrict use to personal and
2:56non-commercial purposes.
2:58uh which means they can't exploit the
3:00service for any commercial purpose
3:02including facilitating advertisements or
3:04solicitations which is a little bit
3:07insane to me because I feel like there's
3:10absolutely no harm in just brainstorming
3:12ad headlines for instance if I'm going
3:15through and I say hey I have this
3:16product can you help me create some ad
3:18headlines for it according to their
3:20terms of services you can't then use
3:22those ad headlines on Facebook or
3:23something like that to help you sell
3:25your product like that is ins insane to
3:28me. Not to mention the fact that any
3:30content that you create with the
3:32intention of selling it or using it in a
3:35commercial manner, like say I wanted to
3:36create a YouTube script that I would
3:38then use on this platform, I couldn't do
3:41that. And of course, if you're using it
3:43to generate pros, you can't do that
3:44either. Quick pause. If you're an author
3:46who wants to start writing more books, I
3:48put together my complete AI prompt pack
3:50that helped me and many others in my
3:52community write a book in a month. It's
3:54completely free. Hit the link in the
3:55description. join the free weight list
3:56for StoryHacker Gold. Then you just
3:58answer a couple quick questions and I'll
4:00send you the full prompt list plus a few
4:02extra free goodies. I'll also email you
4:05when I open the paid Story Hacker Gold
4:07group so you can join if it feels right.
4:09There's no commitment to join though and
4:10you'll still be able to get all of the
4:12freebies you signed up for either way.
4:14All right, back to the video. So, that
4:16was like a huge one for me. And I can't
4:19believe that the pro version, what
4:21you're paying $20 a month for, is
4:23restricted to just like personal
4:25hobbyist type info, right? No. Like, if
4:29you're a business, you're going to be
4:31using this as a productivity tool in
4:32your business. And it doesn't sound like
4:35any usage of that nature would be in
4:38line with their terms of services. Okay.
4:40They also restrict bots or scrapers.
4:42This one doesn't really apply to most of
4:44us.
4:45Um, but it it doesn't really want any
4:48kind of automation software to access or
4:51modify the service itself. Um, not a
4:54huge deal breaker, but something to be
4:57aware of for certain types of
4:58businesses. All right, here's the next
5:00big one. Mandatory attribution. All
5:03published output must clearly site that
5:04it comes from Perplexity's AI services.
5:07No misrepresentation. Users cannot
5:09represent AI generated content as human
5:12created. Uh, that one's fine. In most
5:14cases, especially when the AI generated
5:17content was created with the intent of
5:19sounding real, like for instance with
5:22political parties and stuff like that,
5:24uh any anything political or talking
5:27about like someone said this when in
5:29fact they didn't actually say that. Uh I
5:32understand it for that use when you're
5:33using it as a creative partner. However,
5:36we're creating what is already clearly
5:38fictional content. that to me I don't
5:42know like is is it really AI generated
5:45content when it is a human AI
5:47collaboration I mean yes AI certainly
5:50participated in that whole process but
5:52it also wouldn't have existed without
5:53the human guidance so what's the
5:56difference between that and a human
5:57using any other tool to create something
6:00the tool doesn't create it the human
6:03does with the assistance of the tool and
6:06I don't know that gets into the whole
6:07ethics of the entire debate on on
6:10whether AI generated content uh should
6:12be human attributed or not. Uh so I
6:16won't get into the details of that but
6:18anyway citation required for public use
6:20any public facing or business facing
6:22content using output needs explicit AI
6:24attribution. Um I find it interesting
6:26that even if you know we already
6:28established that you can't use this for
6:30commercial purposes really only for
6:31personal purposes. So even for personal
6:34purposes, anything that you put out
6:36public facing needs to be attributed to
6:38perplexity. I don't know if this is just
6:41them trying to, you know, hedge their
6:44bets against any potential liability
6:46claims or uh or lawsuits that could come
6:49their way or if it's a marketing ploy
6:52like they just want more people to talk
6:53about them using perplexity. I don't
6:55know. Either way, it's a little
6:57annoying. All right. Then we have
6:59professional advice limitations. So, not
7:01for professional decisions. Output
7:03explicitly. Cannot replace medical,
7:04legal, financial, or investment advice.
7:06That makes sense. Uh, no reliable
7:09guarantee. Content is acknowledged as
7:11potentially incorrect, biased, or
7:12incomplete. Users assume all risk.
7:14Strong liability limitations push risk
7:16onto the users for decisions based on
7:18output. And maximum liability capped
7:20perplexities liability limited to
7:22greater to greater of 100 or six months
7:25of payments. So basically, most of this
7:27I'm okay with because yeah, like we
7:30should understand the AI can make
7:32mistakes and we should doublech checkck
7:34anything it tells us uh and not take it
7:37as like it says medical, legal,
7:39financial or investment advice because
7:42or at least not take it at its worst.
7:43You can use it for that advice, but
7:45always double check everything just as
7:47you should if you're googling it. you
7:50know, getting advice from anywhere on
7:51the internet, you should do the same
7:53thing. And acknowledging that it's
7:55potentially incorrect, biased or
7:56incomplete. We know the AI hallucinates.
7:58Even though it's not nearly as big of a
8:00problem as it used to be, it still does
8:02happen. And so, of course, we should be
8:03okay with that. Here's another big one
8:06to keep an eye on. Uh, where it says for
8:09content licensing, user content license,
8:11including inputs, is now explicitly
8:14broad, royalty-free, transferable,
8:15sublicensable worldwide, and irrevocable
8:18while stored. This is one that I could
8:20potentially see becoming a source of
8:22lawsuits for Perplexity if they intend
8:25to enforce this because it means that
8:27anything you put into Perplexity could
8:29be considered basically they have the
8:30rights to use it however they want
8:32including transferring it to other
8:33people. So, while I don't normally have
8:37any sort of fears about putting my
8:39intellectual property into an AI model
8:43because I know that the likelihood of
8:46anybody being able to use that if it's
8:48used as training data, likelihood of
8:50anybody getting to like get ideas from
8:53it or or recreate it in some way is next
8:56to zero. But here it's saying like if
8:58it's in their system, they can use it
9:00however they want. So, with that said, I
9:03would strongly advise you to not put any
9:06of your copyrightable data, any of your
9:10creative writing or anything into
9:11Perplexity because as long as it's
9:13stored there, they can do whatever they
9:14want with it. According to my
9:16interpretation of this, this Reddit user
9:19also um documented a number of instances
9:22of Perplexity banning people from their
9:24Discord or from their Reddit just
9:26because of negative comments that were
9:27said. And you know, while I can only
9:30trust that this Redditor is being honest
9:34about this or has their facts straight,
9:36that does, I think, raise a couple of
9:38questions and a couple of eyebrows in in
9:40uh in that case. But those are the
9:42general red flags that I really wanted
9:44to draw your attention to. Again, I will
9:46put their terms of services and a link
9:47to this Reddit post in the description
9:50below so you can check it out for
9:52yourself. But unless Perplexity changes
9:55some of these things, I am going to
9:57recommend you move to a different
9:59chatbot to do all of your work uh
10:02especially any content related tasks.
10:04The two that come to mind that would be
10:06probably be the best are Gemini and
10:08Claude. Claude does tend to throttle
10:10things a lot and so I tend not to use
10:13that one in particular. I do really like
10:15Gemini these days and just the whole
10:18Google ecosystem makes it one of the
10:21more friendly chat bots I think
10:23especially if you are in the Google
10:25ecosystem at all. If you have Gmail, if
10:27you use Google Maps a lot or Google Docs
10:30in any way, that one might be the one
10:32that I go to for now because Perplexity
10:35is doing all of this stuff. I would
10:37prefer to go to Gemini. I'm still really
10:39bummed out because I would prefer to
10:42have Perplexity's search capabilities
10:44which I think are some of the best.
10:46Gemini has search capabilities as well,
10:48but the other thing that Gemini doesn't
10:50have that Perplexity does is access to
10:53multiple models. The thing about
10:55Perplexity is if I wanted to use a Cloud
10:57model or a Gemini model, I could right
10:59inside of Perplexity. I had access to
11:02about a half dozen or so of the major
11:04models out there and I liked that
11:07flexibility. The problem with ChatGpt,
11:09Cloudp Pro, or Gemini is that you have
11:12to use their models only inside of their
11:15chatbot. You can't use any of the
11:16others. And that was one of the big
11:18selling features of Perplexity for me.
11:20But unfortunately, unless they change
11:22something, I'm going to have to move.
11:24And I recommend all of you do as well.
11:26Hopefully, they make some retractions on
11:28this. A lot of this does smell a little
11:31bit like like a lawyer copy pasted stuff
11:33from other agreements into their terms
11:36of service. So hopefully they can take a
11:39little bit closer look and make sure
11:40that some of these things are smoothed
11:42out and clarified at least because if
11:45they don't, that's a deal breaker for
11:47me. I am gone. So again, I don't make
11:50videos like this too often, but I do
11:54want to make sure that we are bringing
11:55attention to issues like this. And so uh
11:58please reach out to Perplexity if you
12:00can tell them that you're not okay with
12:02this and hopefully we can get some
12:04change done. And with that, I will see
12:06you in the next