Here’s what’s on our plate today:

• 🧸 From Furby to ChatGPT—why AI toys might actually stick this time.

• 🎧 Spotify kills dead air, Meta goes multilingual, and Nvidia builds a chip.

• ✨ 3 A typographic image tool, a voice memo assistant, and a killer AI notepad.

• 📊 Would you buy an AI toy for a kid in your life?

Let’s dive in. No floaties needed…

The Laboratory

From Humane Pin to Playrooms: Tracing the Rise of AI Toys

In 2024, before debates over OpenAI’s GPT-5 dominated headlines, a different conversation was unfolding. At the time, companies were busy promoting devices that they hoped would replace smartphones and help people take their eyes off screens. This was a time of AI devices like the Humane AI Pin and the Rabbit R1. These devices captured imaginations but failed to deliver.

Reviewers labelled the Humane pin “barely reviewable” with buggy performance, poor battery life, hallucination issues, and costly hardware for a function an app could do. The Rabbit R1 and other devices met a similar fate.

Jump to 2025, and AI devices seem to have made a comeback. Albeit with a new target consumer base and a new sales pitch. Promises to adult users have been forgotten, and have been replaced with the new pitch, which looks to sell AI devices wrapped inside toys, which can be an alternative to screen time for kids. However, this pitch has met with more reservations than excitement.

Recently, The New York Times published an article expressing reservations about AI toys. The writer, Amanda Hess, said she felt that the toys with an AI chatbot felt more replacement for a parent, than an upgrade from a teddy bear. She argued that while talking toys might keep kids away from a tablet or TV screen, what they’re communicating is that “the natural endpoint for [children’s] curiosity lies inside their phones.”While this may dissuade some parents from spending on AI toys, it does not seem to be reflective of wider trends.

AI toys are catching on

AI toys (often called “smart toys”) are playthings embedded with artificial intelligence. Unlike traditional toys with fixed responses, AI toys can sense and adapt their behavior based on user input. These toys may come with cameras and microphones to listen to a child’s speech or observe their actions, then use AI software to formulate a response.

Though toys that can respond to a child’s queries have been available in the market for a long time, so far, they have relied on a fixed script. With AI, toys could be venturing into unknown territory. Regardless, the AI toy industry is growing rapidly as the technology advances and consumer interest is building up.

According to a report from Market.us, the market for smart toys, many of which are connected to the internet, was valued at $15.5 billion in 2023 and is projected to reach $77.8 billion by 2033. This represents one of the fastest-growing segments of the $200+ billion global toy industry.

The demand for such toys is being driven by tech-savvy parents, many of whom want to expose their children to the latest technology at an early age. And big toy companies are taking note.

Big toy companies are investing heavily to capture the trend. In June 2025, Barbie-maker Mattel announced a collaboration with OpenAI to bring generative AI to the toy industry. According to a report from Forbes, startups like Miko, which makes robotic companions embedded with in-house AI models along with OpenAI’s GPT-3.5 and GPT-4, have raised more than $50 million and were last valued at about $290 million.

Other companies like Fawn and Curio are also entering the market, showing that entrepreneurs believe they can overcome past pitfalls, of which there are many.

Companies designing AI toys want users to count the benefits. AI toys are screen-free, able to inspire imagination, and make learning more engaging and playtime more enriching. However, beyond the benefits, AI toys come with significant concerns, many of which are backed by past incidents.

The concerns and criticisms of AI toys

The concerns around AI toys range from privacy and security risks to developmental and ethical questions. Child advocacy groups and researchers advise parents to approach these smart toys with caution.

By design, many AI toys are essentially internet-connected listening devices. Conversations recorded and transmitted using such devices are often stored on cloud servers without parents realizing it. In case of unauthorized access to data, these recordings could expose sensitive information about children. Moreover, a lack of strong regulations can also prompt companies to misuse the data. In countries like the U.S., the Children’s Online Privacy Protection Act (COPPA) imposes strict rules on collecting data from under-13 users. However, even companies like Mattel, despite requiring parents to sign a consent and privacy policy, have admitted that they cannot guarantee the security of the data they gather.

In 2017, according to a report from The Guardian, Germany’s telecommunications watchdog ordered parents to destroy or disable a “smart doll” because the toy could be used to illegally spy on children. At the time, the My Friend Cayla doll’s app was found to have flaws that could be leveraged by hackers to directly talk to children.

In another incident, CloudPets, a line of “smart” teddy bears that let kids and parents send voice messages through the toy, was found to have left its online database unsecured.

Apart from privacy concerns, critics point to the ability of generative AI models to produce incorrect or nonsensical answers. These can create problems for users regardless of their age, but for impressionable kids, they can have an outsized impact.

Reliance on continued cloud access and support

Unlike traditional toys, AI toys depend on ongoing support from their makers to function smoothly. Sudden changes in the LLM models being used to power features or a company shutting down operations can impact children much more than adults. User unease around changes in OpenAI’s latest GPT-5 model is a remarkable example of how humans can form bonds with chatbots.

Most recently, the collapse of Embodied (maker of the Moxie robot) highlighted the emotional stakes involved with AI toys. The company announced in late 2024 that it was shutting down and that all cloud services for Moxie would cease. At the time, stories of kids saying goodbye to their robot friend went viral on social media, leaving parents to explain the situation to their children, who might not be ready for the emotional impact. So, while AI toy-makers are ready to reshape the industry, their products have a long way to go before they can earn the trust of parents and their kids.

Caution over wonder

AI toys present a fascinating convergence of play and technology. They have the potential to make toys more interactive, educational, and responsive than ever before. Toys that can talk back with a unique personality were once the stuff of science fiction. And many parents may look at these wondrous inflections of technology and childhood nostalgia as the perfect gift for their kids.

While that may be true, concerns are real. The past is full of incidents where companies made tall promises about the viability of AI devices, be as a replacement for smartphones, or in the form of a robot doll. But that is what they were: tall promises.

As such, AI toys will require balancing innovation with responsibility. Sophisticated toys, especially ones powered by AI, will have to be designed with a kids-first approach in mind. This would entail bulletproofing data security, building strict content filters, and being transparent about how these toys work. As for parents, they must look at AI toys as fun and educational gadgets, not an alternative for active care and attention.

Regardless, AI toys are here real whether we like it or not. Whether they fizzle out like the Humane Pin or cement their place in the mainstream with backing from companies like Mattel and Hasbro remains to be seen.

Quick Bits, No Fluff

• Spotify smooths the vibe: The streaming giant is testing seamless audio transitions between playlist tracks, aiming to kill the dead air and keep the flow going.

• Meta brings global AI flair: Creators can now access AI-powered translations starting with English ↔ Spanish, as Meta expands its multilingual push.

• Nvidia builds a new chip for China: The company is reportedly developing an AI chip that outperforms its H20 model—without violating U.S. export rules.

Thursday Poll

3 Things Worth Trying

• Ideogram: AI image generator that nails typography — perfect for poster-style visuals.

• Voicenotes: Turns your voice memos into structured notes, summaries, and tasks instantly.

• Granola: AI notepad for people in back-to-back meetings. Turns meeting notes and makes them awesome.

Rate This Edition