More than 36 million child-friendly AI and game credentials compromised in 3 years

Millions of Roblox IDs and popular games are circulating on the dark web, putting children’s safety at risk.

Digital security is more than ever at the heart of concerns with the recent publication of the Kaspersky report, revealing an astronomical number of 34 million Roblox credentials stolen on the dark web since 2021, seriously compromising the safety of children online.

However, this threat is not limited to online gaming, as indicated by an alarming 112% increase in compromised accounts on major platforms such as Twitch, Electronic Arts, Sony PlayStation and Steam, or artificial intelligence applications , thus highlighting the growing scale of cybercrime in the digital entertainment sector.

Children, preferred target of cybercriminals

According to Kaspersky’s report, nearly 34 million Roblox credentials have been stolen and circulated on the dark web since 2021. These credentials mainly belong to children, who are more vulnerable to data-stealing malware. Cybercriminals use social engineering techniques, such as files containing cheat codes, to infect young gamers’ devices and gain access to their accounts.

“ For example, cybercriminals can hide information thieves in files containing cheat codes to deceive young gamers. In some cases, this deception may appear genuine, as malicious download links may be posted on legitimate and popular social media platforms like YouTube said Yuliya Novikova, head of Kaspersky Digital Footprint Intelligence.

Furthermore, the average number of compromised accounts on various gaming platforms increased significantly in 2023. Despite the large volume of Roblox IDs available on the dark web, the report reveals a marked preference by cybercriminals for Steam IDs. This preference is explained by the possibility of finding and stealing real money from Steam accounts, in addition to valuable items such as expensive skins. As Yuliya Novikova explains, “ Steam accounts thus become more attractive targets », adding a financial dimension to cybercriminal activities.

Growing demand for compromised AI accounts

A worrying trend is emerging with a dramatic increase in stolen credentials for OpenAI services, including ChatGPT, recording 664,000 cases in 2023. The release of ChatGPT 4 in March 2023 caused a spike in demand, a report confirms later Check Point.

This increase in activity on the dark web highlights the vulnerability of artificial intelligence services to cybercrime, putting user privacy at risk.

Besides ChatGPT, Kaspersky reveals that AI-powered services, such as Canva and Grammarly, are not spared, with more than 1 million Canva user IDs and approximately 839,000 Grammarly user IDs compromised over the course of of the last three years. This trend highlights the diversification of cybercriminals’ targets towards services that are increasingly integrated into our digital daily lives.

Cybercriminals can use compromised AI accounts to manage automated bots for spreading spam, disinformation or manipulating financial markets, or to access sensitive information such as personal data or user training data or
hijack AI services for malicious purposes.

Source : Kaspersky