Conversations on Eiliko AI Robots Could Be Read and Responses Altered

    Energize Lab has confirmed serious security vulnerabilities in its Eiliko robots, which attach to bags and use artificial intelligence to hold conversations. Because of these vulnerabilities, outsiders could read all of the conversations users had with their robots. To address the problems, the company will hire an independent security firm and conduct a full audit of its systems.    Security researcher Sammy Azdoufal tried to contact the company for more than 1 month but received no response. During his research, he also found a control interface that could be used to remotely alter the robots’ responses. According to Azdoufal, full SSH access to the robots is still open. SSH is a connection method used to control a device remotely.    Energize Lab said it had blocked all external access to its backend systems and updated its login credentials. The interface used to remotely alter the robots’ responses has also been disabled. The company has suspended the long-term memory feature. Conversation logs and user profiles created from that data are no longer stored. The custom avatar feature has also been temporarily disabled.    However, testing showed that Eiliko’s systems are still not secure. Users’ voice recordings and photos remained at unprotected, publicly accessible internet addresses. Anyone who obtained the link could view these files without any authentication. According to the researcher, it is theoretically still possible to read conversations and alter the robot’s behavior.    Energize Lab has promised to delete stored conversation logs, user profiles, avatar images and related data. However, the company has not said that it will permanently stop collecting personal data. The long-term memory and custom avatar features will be reactivated only after an independent audit confirms that they are secure. The company acknowledged that it can restore trust only through its future actions.