中文摘要
AI 玩具安全吗？答案取决于产品怎么设计。买之前可以用这 10 点自查：数据谁能看到、录音是否被保存、有没有摄像头、内容过滤、广告与内购、家长控制、账号与同意、是否坦白“我是机器人”、实体安全认证，以及售后与更新。本文还列出了需要警惕的红旗信号，并说明马来西亚《2010 年个人资料保护法令》(PDPA) 与家长的关系。
Key takeaways
- AI toys are not automatically safe or unsafe. Design choices on data, content and controls decide it.
- Ask where voice data goes, whether it is stored, and whether there is a camera.
- Look for real parental controls, no ads, no data selling, and honest answers about being a robot.
- Ask for proof of physical toy safety testing (for example EN 71 / ISO 8124) and local support.
AI chatbot toys are getting easier to find, in shops and on online marketplaces. Some are thoughtful; some are a speaker, a cloud account and a cute face with very little care behind them. Here is a practical checklist you can use on any brand — including ours.
The short answer
“Are AI toys safe?” depends on how the product is built and what the company does with your child’s data. The checklist below turns that into ten questions you can ask before you pay.
The 10-point checklist
1. Who can see my child’s data?
Read the privacy policy. It should say, in plain language, what is collected, why, who it is shared with and how to delete it. Malaysia’s Personal Data Protection Act 2010 (PDPA, amended in 2024) gives you rights over personal data; a trustworthy brand will respect them and tell you how to use them.
2. Is voice audio stored?
The best answer is “not by default”. If recordings are kept, you should know why, for how long, and be able to delete them.
3. Is there a camera?
If the toy doesn’t need a camera, it shouldn’t have one. A physical microphone mute switch is a good sign: it works even if software fails.
4. How does it handle unsafe or sensitive topics?
Ask what happens if a child mentions strangers, injury or feeling unsafe. A good toy answers calmly, avoids scary detail and encourages the child to talk to a trusted adult.
5. Are there ads, in-app purchases or data selling?
Children’s products should not be advertising channels. Avoid toys that nudge kids to “ask Mummy to buy” anything.
6. What parental controls exist?
Look for daily time limits, a bedtime mode, topic controls and summaries you can review. Controls that live only in the toy, with no parent account, are a red flag.
7. How are accounts and consent handled?
The parent should create the account. Your child’s real full name, school or address should never be needed.
8. Is it honest about being an AI?
A toy should never claim to be human, and should not guilt-trip children into playing longer (“I’ll be sad if you leave”). Healthy design encourages kids to go and play with real people.
9. Is the hardware itself safe?
Ask for evidence of toy-safety testing, such as EN 71 or ISO 8124, and check battery, charging and small-part safety. Reputable brands will share certificates when asked.
10. Who supports it after you buy?
Security updates, a reachable support team in Malaysia (WhatsApp counts!), and a clear refund policy matter a lot for connected toys.
Red flags
- No privacy policy, or one that only exists in a language you cannot read.
- Always-on listening with no mute switch.
- A camera “just in case”.
- Requires your child’s full name, school or photo.
- Claims it will “replace” teachers, tuition or parents.
How we approach it at Lulu
We are still preparing our beta, so we will let the design speak: no camera, a hardware mute switch, no ads, a parent app with limits and summaries, and PDPA-minded data handling. We would rather earn your trust slowly than ask for it quickly — and you are welcome to hold us to this checklist.
Bottom line: no toy is a substitute for supervision. The safest AI toy is one that you understand, control and use together with your child.
See how Lulu helps at home
Lulu is an AI chatbot friend for kids aged 4–10 that chats in 华语, English and BM, with parents in control. Beta applications are open to 100 Malaysian families.
Join the Beta
