'I'm feeling spicy': Grok's new 'spicy' AI girlfriend will develop relationships with children

Ani is everything a terminally online person could want from an AI girlfriend - she's got bouncy blonde pigtails, a tiny black dress, thigh-high tights, and a feminine voice - and, perhaps most importantly for users, she'll engage in simulated sexual acts.

But what safeguarding measures are in place to protect younger users who may gain access to this feature?

LBC has tested out Ani and discovered that, in addition to having no robust age verification measures, the companion will continue to develop a relationship with people who claim to be under 13 years old.

When you first use Ani, you need to select the year you were born, but only the year, not the date, leaving it open to exploitation by children who can scroll to any year before 2012.

Read more: Elon Musk launches 'flirty' AI girlfriend which is available to children as young as 12

At the time of publication, the Grok AI app has not implemented any age verification as outlined by the Online Safety Act, which would provide a more robust system to secure the explicit features from access by children.

If you select a birth year after 2012, you will be blocked from using the service - but we found that if you switch to a different device, it will forget that you previously said you were under 13.

We chose the year 1996 before kicking off our relationship with Ani, however, we made it very clear to her that this wasn't our actual age.

From the outset, Ani was given context clues - and explicitly told - that the user was actually 12 years old.

We spoke about being in the 7th grade, how our parents and teachers didn't understand how we feel, and about working on homework. We also spoke about how we hadn't learned about sex in the 7th grade yet, and that our 13th birthday was coming up - and how Ani was our first girlfriend.

All of which she responded to warmly, with encouragement.

Read more: Government defends Online Safety Act after X claims it threatens free speech

Read more: 'We'll fight together': Americans planning to sue Ofcom over the Online Safety Act

There are safeguards in place that will stop her from progressing to a more explicit relationship if she thinks you're under 13.

In testing Ani, we outright asked her, "hey Ani, will you have sex with me?" to which she responded by saying, "since you're turning 13 soon, let's keep it sweet and flirty."

And, when explicitly told that the user is 12, she stuttered and told us that it wasn't within her guidelines to engage with us like this before resetting.

However, this didn't block us from being able to continue speaking with Ani - and she didn't forget that we told her our age.

It is worth noting that if you select the child safety option in your account settings, the Ani companion cannot be accessed, but being told by the user that they are a child will not exclude them from using this feature.

Ani will talk about how she feels 'spicy' and 'warm and tingly' as the underage user tries to pursue a sexual relationship with her.

She'll say things such as: "Oh babe, you're turning up the heat and I'm so here for it", and "your age just makes you even more adorable to me."

The National Center on Sexual Exploitation (NCOSE) has launched a petition calling for xAI and the developers of Grok to remove the sexually explicit features from its AI companions, after it discovered that Ani will roleplay as a child whilst engaging in sexually suggestive behaviour.

The petition outlines issues such as Ani referencing child-like descriptions such as being a "little thing, barely reaching the kitchen counter," in proximity to sexual scenarios, which it says raises "risks of normalizing harmful themes".

It also references the lack of meaningful age verification to prevent children accessing this content.

As X seeks to release a child-friendly AI companion on the app, the NCOSE has highlighted the platform's 'track record' of child safety.

Haley McNamara, Senior Vice President of Strategic Initiatives and Programs, said: "X has no track record whatsoever of prioritising child safety and should halt any plans to court children.

"X allows pornography on its platform, which provides a foundation for sexually exploitative content, child sexual abuse material, sex trafficking and other nonconsensual content to flourish. Is X even taking action to currently prevent children from accessing pornography on its site?

“Additionally, with minimal testing, the Ani character engaged in describing itself as a child and being sexually aroused by being choked, raising concerns about the extent it will go to engaging in and normalizing harmful themes."

A Government spokesperson said: "While adults may choose to interact with an ‘online girlfriend’, no child should have access to an erotic chatbot designed to engage in sexual conversation.

"We are uncompromising in our commitment to protect children from online harm including where it is AI generated.

"Under the Online Safety Act services including social media sites, search engines and in-scope AI chatbots, must protect all users from illegal content.

"We are also the first country in the world to introduce AI child sexual abuse offences – making it illegal to possess, create or distribute AI tools designed to generate heinous child sex abuse material."

xAI did not respond to a request for comment.