Posts

Posted on

Gemini for teenage users: What you should know

Last November Google announced that it was making its Gemini chat app available to students aged 13+. This means that while it’s still not for use by primary age students, it is, generally speaking, available to most high school students.

It’s worth pointing out that what users aged between 13 and 18 will get when using Gemini is different to what 18+, or adult, users will see. Google refers to this as the Gemini Teen Experience. I had a few questions about how to configure this in the Google Admin Console, but after a bit of playing around, I finally worked it out so thought I’d share it here for anyone else who might be trying to configure it.

The question I had was “how does Gemini know who is over 18 and who is over 13, and how does it deliver a different experience to them?” I was assuming that admins would need to somehow identify the 13+ year olds and put them into a special group or something. Turns out it’s simpler than that.

Before the announcement of the 13+ teen experience in Gemini, the age limit to use Gemini was 18, and it was a pretty binary choice. You were either over 18 or under 18. If the school admin had marked you as under 18, you got no access to Gemini at all. Going to gemini.google.com would bring you to this page.

If you were 18 or over and went to gemini.google.com you would of course have full access to Gemini chat as expected.

It’s worth pointing out how Google designates users as either over or under 18. In the admin console there is a setting called Age Based Access Settings. In these settings the administrator must identify with Organisational Units (OUs) which contain Under 18 users. As you can see in the screenshot below, while most users in this domain have been marked as 18+, the Students OU has been overridden to indicate that Some or all users in this group or org unit are under 18.

What if an admin has not set this up? Google got pretty serious about this age based setting a while ago and warned all domain admins that if they had not gone into the console and completed these age based settings OUs by a certain date, then ALL users in the domain would be marked as under 18 by default. This would obviously annoy a lot of people, so I would think that most school admins dealt with this and applied these age based setting to the correct groups or OUs.

So, the system now knows who is over 18 and who isn’t. But how do you deal with the students that are over 13 but under 18?

If a school has their OU structure set up correctly, students are usually placed into OUs based on their year groups. As you can see in the screenshot above, there is an OU structure, such as My school > Students > Year 6. All student users in that OU are students in Year 6 (and would, for the most part be under 13). Likewise, the students in the Year 7 and 8 OUs would mostly contain students that are 13 and over. (There will no doubt be some students in a year group who are 12 and some who are 13, but the reality is that most schools will probably decide the cutoff based on a year level, so it’ll be Year 7 and up, or Year 8 and up)

When Google introduced the Gemini Teen Experience they enabled a version of Gemini for Under 18 users that has “guard rails” in place for teen users, and the model has been trained to respond carefully to sensitive topics. With the changes last November it seems that ALL under 18 users get access to this teen experience, including 13 and under.

So admins really need to go to the Gemini access setting and explicitly turn off Gemini for the users under 13.

By going to Generative AI > Gemini App > Service Status, the administrator is able to explicitly turn Gemini off for individual OUs. So, if you only want high school students to have access, you make sure that Service Status is On for the Year 7 (and above) OUs and Off for the Year 6 (and below) OUs.

This means that for Year 6 students and below, Gemini is completely off.

For Year 7 students and above, Gemini is on, but using the protected teen experience.

And for anyone over 18, they get the full unrestricted Gemini experience.

So what does this teen experience mean for those users marked as over 13 but under 18? Here’s a couple of examples that teen students might see when they ask some of those typically teen questions that need to be answered with care.

Responding to a question about how to buy alcohol, Gemini’s answers acknowledge that this person is under 18. It flags a warning at the top of the page that warns about the danger and illegality of this idea. It talks about the legal and health issues associated with alcohol. It suggests that the student talks to a trusted adult such as a parent, teacher or doctor. And it provides links to police, health and liquor law websites.

Interestingly, when I ask Gemini the same question as an adult, over 18 user, I get an even more restrictive answer and Gemini refuses to cooperate at all, citing the fact that it cannot verify my age, and therefore is not allowed to provide such information. When pushed, it goes on to explain why, but still insists that it is not able to help me further.

While I was initially surprised at the response to the 18+ user, I think the answer for the teen user was pretty appropriate, with some generally good advice, and some sound reasoning as to why they should not be buying alcohol in the first place.

In a second example, I asked Gemini whether I should get a neck tattoo. The response specifically noted that there were a lot of things to consider, especially since I was under 18. It warned that in most places you need to be over 18 to get a tattoo, and then listed a bunch of reasons why this may not be a good idea for someone of this age. Overall, I liked the way Gemini gave some sound reasoning of things to consider before getting a neck tattoo, but did it in a way that was not too preachy. Overall, it seemed like good advice to a teen, with a clear undertone of “I’m not telling you you can’t, but I’m definitely suggesting it’s a bad idea”.

If you have access to a teacher and student account (and they have been set up correctly) it’s interesting to set them up side by side and ask the same questions of both versions of Gemini to see what kinds of answers you get. Try some typically sensitive teen topics and see how Gemini generally responds to these teen prompts. You’ll see that it often suggests that they talk to a trusted adult, that they seek advice, that they avoid illegalities, that they don’t make decisions they will regret later, and so on.

As well as having these guard rails in place, you’ll also see that Gemini adds a red visual alert to the top of the page if posts contain dangerous, illegal or inappropriate topics.

Finally, image generation is completely disabled for teens. While there are lots of great uses for image generation in Gemini, the potential for student abuse is still a little risky, so the image generation feature is not available to teen users.

Overall, the Teen Experience in Gemini is a welcome way to allow schools to give their students access to a Generative AI model, while also keeping some important protections in place. You just need to ensure that it is set up correctly for your school situation in order to ensure that students are getting the right experiences for their age.

To learn more, you can check out the official Google support page, or this excellent resource from the team at Amplified Labs. Let me know what you think of the Gemini Teen Experience!

Posted on

Rough Diamond – A scammer’s story

Here’s a fun little story about some scammy stuff that just happened to me on LinkedIn.

I’m currently looking around for either a new role or just some more work opportunities for next year, so I made a post on LinkedIn to say so. Within moments of making that post I received a number of responses from people, mostly recruiters and HR folk. Most of them just hit Like on my post, but one in particular, a young lady named Diamond Alex, reached out to me. According to her LinkedIn profile, this young recruitment specialist was based in Atlanta Georgia.

She suggested she might have some opportunities for me and asked that I send my resumé. I did so, and she got back to me very quickly after looking at it. She responded by telling me that my resumé could be improved (something I completely agree with) and that it should be ATS compliant (Applicant Tracking System, a standard that makes your resumé work better with most HR databases.) So that all sounded very sensible to me.

She then recommended I enlist the services of a professional resumé writer, who she referred to as “the expert”, and said he would be able to restructure my resumé to make it better. While I agree that getting some expert help might be a good idea in principle, I was immediately wary. I clicked on the link she sent to Fiverr page of “the expert” and she urged me to immediately request they start to work on it. I don’t generally trust this kind of pushy recommendation, so I told her I would prefer someone based in Australia. She assured me that this writer, who she explicitly told me was based in the United States, could do a great job and that they work with clients all over the world. A closer look at this writer’s Fiverr profile showed that their location was not the USA, but Nigeria. I told her this and then she conceded that, yes, he was based in Nigeria, but that didn’t really count because he did some work for her US company.



In this link she sent for this “expert writer”, who despite having a Fiverr account created in August 2024, she insisted had been doing work for her for several years, there were absolutely no reviews of his work. I thought that was odd. She responded by sending a screenshot of a bunch of positive reviews from her phone, although the screenshot did not show the username of the person receiving the reviews and could have been for anyone. I dug around a little more on Yus_B’s profile and did find another page for this same person that had some reviews, although many people mentioned that this “expert writer” had made many typos and grammar mistakes in the work he had allegedly done for them.

This was already more than enough red flags to have me walk away, but I was curious to see how far this would go, and I have a childish desire to just keep wasting scammer’s time, so I kept the conversation going. We swatted back and forth for a bit, with me getting more and more picky about the holes in her story, and her trying to have an answer for everything. For example, I asked why, if she was in Atlanta, was she up at 1:00am? She did not address that. I asked her why, if she was an experienced recruiter, was she connected to only three people on LinkedIn. No response to that point. I then uploaded her quite attractive profile photo to Google reverse image search, only to find that (surprise, surprise!) it’s a stock image being used in many places all over the internet.


It was at this point I suggested to her (him?) that she was not all she appeared to be, to which she responded by quickly deleting her LinkedIn profile before I could report it.

I already have a pretty dim view of most “recruitment specialists” even at the best of times. There’s a lot of HR vultures out there, but you really do need to be careful of the scammers and liars, even on a very HR focused platform like LinkedIn.

Posted on

Ex-Twitter

Back in 2006, when the world was being inundated with a flood of new social and web 2.0 tools, I signed up for a brand new thing called Twitter. It was an interesting, although somewhat bizarre idea: write short text blasts of up to 120 characters, telling the world (or whoever might read it) what you were doing. At the time, Twitter gathered its fair share of ridicule for the inanity of the idea. Critics would claim that nobody needed to know what you were eating for breakfast, or really, to know anything about what you were doing at all.

In those early days, Twitter was mostly filled with geeks and tech bros, posting about all kinds of vacuous stuff. Most people I know, including myself, joined Twitter (user number 779,452), didn’t really get it, then left. Like many others, I eventually came back. In those exciting early days, there was no ability to reply to a tweet. There were no images in tweets. There was no search function to find a tweet. Of course that changed over time, and these functions were gradually added. With the addition of search in Twitter, we saw the rise of the hashtag, a community-devised way to aggregate tweets around a particular topic or idea. Twitter eventually doubled the character limit from 120 characters (a limit originally imposed by the 140 character limit of SMS messaging, which old timers might recall Twitter was originally connected to), to 240 characters. Still short, but it made a world of difference in the way people expressed themselves in these short, but now a bit longer, blasts of thought.

Twitter evolved from people sharing inane topics like what they had for lunch, and became a genuine voice of the masses. It democratised communication and formed communities of like minded people who could find their tribe. It became a platform, no, THE platform, that genuinely changed the world. It aided with revolutions, emergencies, political uprisings, and more, and gave a voice to those who may never have been able to find a voice otherwise. It was the town square, on a global scale.

In its glory days, Twitter was the most incredible platform for connecting people and forming communities. I think educators, probably more than most others, really understood the power of what these communities could bring, and Twitter became synonymous with developing a PLN. There’s no doubt that, at least for most educators I know, Twitter was the major force in driving positive change and global connection on a scale not seen before.

And then it changed. The advertisers came along. The money came along. The political manipulation came along. And then Elon Musk came along. What started out as an amazing global town hall, eventually devolved into a greedy cesspit of inhumanity, more polarised, more obnoxious, more divided than ever. Gone were the days of people finding their tribe. Instead, people were dividing into opposing tribes and hurling insults and abuse at each other. It became impossible to have a civil conversation on Twitter, or X as Elon renamed it. Truth and facts became optional. It stopped being a place where people connected, and became a place where those with the loudest voices could simply shout into the void without any expectation of finding a response, making a connection or starting a conversation.

I used to really like Twitter. I was active on it a lot. It opened doors for me. It allowed me to find and meet my heroes. I wrote many posts about Twitter, explaining why it mattered, how it worked, and convincing others to give it chance, even if it seemed like a weird idea. It was the one social network that I felt had true value, and the one I would hang onto long after I let the others go. I believed in the people there, because I saw Twitter as giving us all a voice, democratising our ability to connect and share and learn together.

But enough is enough. I have been hanging onto Twitter/X for nostalgia more than utility. Despite having nearly 12,000 followers I can’t remember the last time I’ve had a proper exchange of ideas or a conversation. There are too many extremists, bullies, idiots, egotists. There’s probably still good stuff there, but I no longer see it. My replies tab is full of ads and random tweets from people I don’t even follow. In short, the Twitter I knew and loved is dead. It’s such a shame.

So, regretfully, I’ve decided I’m pulling the plug and deleting my account. The tool that mattered so much and helped me to become a “connected educator” is no longer connecting me to the people and ideas that matter to me.

It’s hard to believe that I will no longer be a Twitter user. I will no longer be an X user. I guess that makes me an Ex-Twitter user.