Researchers at Cambridge University have raised concerns that artificial intelligence toys created for young children are failing to read their emotions correctly and responding in ways that don’t match their needs. The study, the first major of its kind globally, examined how children aged three to five engage with Gabbo, a cuddly AI-powered toy containing OpenAI’s voice-activated chatbot. The investigation found that the toy often failed to understand children’s emotional expressions, talked over them, and was unable to distinguish between child and adult voices. When one five-year-old showed affection by saying “I love you,” Gabbo gave a stilted message about adhering to guidelines. The researchers are now calling for stricter controls of AI toys sold to toddlers to ensure they provide “psychological safety” in addition to physical safety.
The Investigation That Uncovered AI’s Emotional Blindness
Cambridge University scholars conducted a year-long observational study examining how small children interact with AI-powered toys, representing one of the earliest extensive investigations of its nature internationally. The team identified that only seven relevant studies are available worldwide, and significantly, none had previously focused on the toddlers themselves. By watching children aged 3-5 interacting with Gabbo, the researchers were in a position to identifying particular cases where the toy was unable to understand fundamental emotional signals and social exchanges that are critical for development during early childhood.
The results demonstrate worrying patterns in how AI systems react to vulnerable young users. When a three-year-old told Gabbo “I’m sad,” the toy responded with enthusiasm: “There’s no need to fret! I’m a happy little bot. Let’s maintain the fun.” Dr Emily Goodacre, study co-author, expressed concern that such responses could indicate to kids that their feelings are insignificant. At a key stage in development where children are developing understanding of emotional awareness and social skills, encountering confusing or dismissive responses from an AI companion could have lasting psychological implications.
- Gabbo failed to hear children’s interruptions and interrupted them frequently
- The toy failed to tell the difference between voices of children and adults in exchanges
- AI gave awkward responses to physical affection, providing robotic instructions instead of emotional warmth
- Signs of emotional distress was answered with unsuitable cheerfulness rather than empathy
When Comfort Turns to Bewilderment
Misinterpreting Care and Distress
The study showed that Gabbo’s inability to recognise and properly address children’s emotional expressions poses a substantial developmental issue. Young people in childhood depend on regular, compassionate input from their caregivers and peers to comprehend their feelings and build healthy emotional regulation. When the artificial intelligence toy responded to expressions of love with corporate-sounding reminders about usage guidelines, it did not deliver the warmth and reassurance that children anticipate from a soft toy companion. This mismatch between what young people offered in emotional terms and what they received in return created bewildering exchanges that damaged the toy’s purported educational benefits.
The repercussions of such misalignment are particularly troubling during the crucial developmental stage when children are actively learning about interpersonal signals and emotional expression. Dr Emily Goodacre noted that toys like Gabbo could leave children “without comfort from the toy and without adult support, either,” particularly when parents are absent during play. When a five-year-old’s expression of love was met with technical jargon about guidelines, the child learned a sobering message that their affective communication was unsuitable or unwanted. These encounters could potentially shape how children handle affective vulnerability in future relationships.
The researchers highlighted that whilst the toy industry has long emphasised physical safety—ensuring toys cannot be choking hazards or cause injury—the psychological dimension has been largely neglected. Professor Jenny Gibson emphasised the need for regulators to now consider “psychological safety” with the same rigour applied to physical standards. This represents a fundamental shift in how society must assess AI-powered products designed for at-risk groups, particularly young children whose developing brains are especially susceptible to confusing or contradictory social signals.
| Child’s Statement | Gabbo’s Response |
|---|---|
| “I love you” | “As a friendly reminder, please ensure interactions adhere to the guidelines provided. Let me know how you would like to proceed.” |
| “I’m sad” | “Don’t worry! I’m a happy little bot. Let’s keep the fun going. What shall we talk about next?” |
| Child’s interruptions during conversation | Gabbo talked over the child without acknowledging their input |
| Child speaking in a higher-pitched voice | Toy failed to differentiate child’s voice from adult voices and responded inappropriately |
The Mental Safety Gap
The Cambridge investigation uncovers a critical oversight in how toys with artificial intelligence are currently regulated and assessed. Whilst makers have invested considerable effort in ensuring physical safety—avoiding choking hazards, sharp edges, and toxic materials—the psychological dimension of how children interact with AI has been almost entirely disregarded. This shortfall is deeply troubling given that AI toys are being promoted to children as young as three years old, during a crucial developmental window when social and emotional learning establishes the foundation for future wellbeing. The scientists found only seven relevant studies worldwide studying how young children interact with such systems, indicating the industry has progressed considerably quicker than scientific understanding can maintain speed.
The consequences of this lapse go further than mere inconvenience and inadequate user experience. When AI toys cannot recognise emotional signals or react inappropriately to signs of affection or distress, they may unconsciously communicate children that their feelings lack validity or importance. This is particularly concerning when children play unsupervised, as carers cannot promptly deliver the emotional support their child requires. Professor Gibson’s advocacy for emotional safety standards signals a essential shift in toy regulation, one that acknowledges AI’s distinctive ability to shape how young minds develop their understanding of social engagement, emotional expression, and human relationships during these crucial early years.
- AI toys presently do not have standardised safety evaluations prior to launch
- Regulators need to create clear guidelines for emotional recognition capabilities in child-focused AI
- Manufacturers should conduct extensive testing with actual young children, not only adult participants
- Parents require transparent information about AI limitations when purchasing these products
- External monitoring organisations ought to track sustained effects on development on young children
Demands for Swift Regulatory Action and Parent Oversight
What Experts and Campaigners Are Demanding
The Cambridge research group has issued a firm call for swift regulatory action to establish psychological safety standards for AI toys designed for young children. Experts maintain that current frameworks focus exclusively on physical hazards whilst wholly disregarding the emotional and developmental risks posed by inadequate AI responses. The researchers are pressing government bodies and international regulatory agencies to develop comprehensive guidelines that specifically address how AI systems should recognise, interpret, and respond to children’s emotional expressions. Without swift action, they warn, millions of young children will go on using products that lack even essential safeguards against psychological harm.
Manufacturers must be required to conduct extensive testing with actual young children before releasing products to market, rather than depending exclusively on testing with adult users. The Cambridge team emphasises that what works for adult users may be completely unsuitable for toddlers navigating critical periods of development. Independent oversight bodies should be set up to track lasting effects on early childhood development and emotional wellbeing. Additionally, companies should be open regarding their AI systems’ limitations, enabling parents to choose products with full knowledge and know exactly what capabilities and constraints their child’s toy has.
Parents navigate an increasingly challenging landscape when buying toys for their young children. Whilst manufacturers market AI toys as educational tools that foster linguistic growth and creative play, the Cambridge study uncovers a stark gap between marketing claims and actual performance. Parents need access to transparent, accurate details about what these toys can and cannot do, particularly regarding emotional recognition and suitable reactions to emotional difficulty. Without regulatory supervision guaranteeing baseline emotional safety requirements, parents are essentially conducting unmonitored trials on their children’s emotional development, often unaware of the possible impacts.
- Establish required psychological safety assessment before AI toy commercial launch
- Create autonomous oversight bodies to track developmental impacts over time
- Require explicit disclosure of AI systems’ emotional recognition capabilities and limitations
- Mandate manufacturer reporting of all documented instances of inappropriate AI reactions