Miko, a company that makes artificial intelligence-powered toys, is incorporating an AI on-off switch into its core technology following political pressure and product scrutiny.
The new option, which follows public criticism, is a rare adjustment that allows consumers to step back from using AI at a time when companies in all sectors are rapidly integrating the technology into their products.
“Miko puts complete control in the hands of parents and caregivers with an on/off toggle option to enable or disable conversational AI features on Miko 3 and Miko Mini,” the company announced in a press release on Monday.
Miko toys are animated robots with touchscreen faces that can play music and games with children and interact with them using an AI chatbot called a Large Language Model (LLM). According to a 2024 blog post, the company worked with Google to use Google Cloud and Google’s Gemini AI model.
Miko has recently come under intense scrutiny and political pressure, with watchdogs and politicians raising safety concerns about Miko and other AI toy makers.
Last week, Sens. Marsha Blackburn, R-Tennessee, and Richard Blumenthal, R-Connecticut, said they had identified a website where any visitor could download thousands of responses Miko had created for specific children. The file appears to have been accidentally left online without password protection or other guardrails in place, making it accessible to anyone with an internet connection.
NBC News heard part of the recording. The recordings included what appeared to be a lengthy conversation in which LLM repeatedly called the child by name and asked questions such as how she was feeling and what kind of music she was listening to. Last week I was unable to access the site.
Miko CEO Sune Vaswani said in an emailed statement to NBC News at the time that the company did not leak user data or store audio recordings of children, but did not directly comment on the recordings that appeared to be responses to children.
Blackburn said in a statement following the news of the AI off switch: “These new parental controls are an 11th-hour attempt to save face following the cybersecurity breach revealed last week in which the company released sensitive data about children to the public. Regardless of the latest press release, parents should think twice before purchasing this toy for their children.”
Ritvik Sharma, Miko’s senior vice president of growth, declined to respond directly to Blackburn, but said the new feature “has been in development for some time” and that the audio response posted online “has nothing to do with us making that announcement.”
AI toys and the chatbots that power them are largely unregulated in the United States. The popularity of toys has skyrocketed recently as technology becomes more widespread and popular.
Toy manufacturers usually have safeguards in place to avoid saying inappropriate things to children. However, all LLMs are susceptible to “jailbreaking,” a phrase that causes the chatbot to ignore the developer’s instructions.
In December, an NBC News investigation found that some of the most widely sold AI toys to Americans could, in some situations, make them say things that parents might find worrying or uncomfortable, such as depictions of sexual acts or explanations of geopolitics closely aligned with the Chinese Communist Party.
The AI toy market is growing rapidly, especially in China. MIT Technology Review reported last year that the company has more than 1,500 registered AI toy companies.
