Kaspersky Gives Advice on How to Make AI for Children Safer @ Safer Internet Day

Born between 2010 and 2025, Gen Alpha aren’t just growing up with technology – they’re actively living it. These digital natives are already wielding smartphones, tablets, and AI-powered tools with the confidence of seasoned users, navigating everything from gaming and social media to online learning platforms with remarkable ease.

But the question that concerns parents and security experts is whether we are giving our children too powerful technology, too soon. On Safer Internet Day, Kaspersky security experts are sharing practical tips to help parents turn AI from a potential threat into a trusted ally for the younger generation.

The first line of defence is building AI awareness

Children have already discovered that ChatGPT, DeepSeek and other neural networks can answer questions faster than you can find the right answer in Google, and Alexa can play music without pressing a single button.

So, the only solution is to become children’s AI support. Begin by explaining that these digital assistants aren’t friends, pets, or even real people. They’re sophisticated tools that can be helpful, but also potentially misleading, biased, or simply wrong. Then teach them to cross-check information with multiple sources, just like they’d verify facts in a school project.

When discussing AI with children, emphasise that they should never fully trust AI answers, especially for sensitive topics like health, mental wellbeing, or safety concerns. Always encourage them to verify information and never share personal details or documents with AI systems.

Enabling safety filters

Most AI platforms and smart devices come with built-in safety features that are often overlooked or misunderstood. Spend some time checking the privacy settings and content filters and, if possible, tailor them to match your family’s values and your child’s maturity level. This is a basic protection against inappropriate content, privacy breaches, and potentially harmful interactions.

However, not all services and platforms provide an opportunity to set up content filters and fully control children’s online activity. To create a safer digital environment for your children consider using parental control tools like Kaspersky Safe Kids. It allows parents to not only hide inappropriate content and prevent specific apps and websites from being opened but also helps balance children’s time spent online with screen time management.

Checking the AI-powered apps authenticity

In a world where AI apps are popping up faster than you can say “chatbot,” verifying app authenticity is essential. Only download apps from official stores and inform your children about the importance of not installing anything from unfamiliar sources. Look up the company behind the app and check whether they have a website and legitimate business presence.  Teach your kids to limit their app’s permissions and do not give access to data unless it’s necessary for the apps to work.

Staying involved and informed

A basic understanding of the range of problems your child is willing to entrust to AI is already significant. By asking simple questions like “What did you ask AI today? Did it give you the right answer?” you’ll be teaching your children to openly discuss with you the use of AI and problems they might face. When they mention using ChatGPT for homework, ask them to show you what they’ve learned. When they talk about their favourite voice assistant, ask about the topics they like to discuss and funny particularities they noted.

“When you actively participate in your child’s AI journey, you transform from a concerned parent into a trusted guide. They’ll seek your input because they know you’re interested in their digital experiences, not just trying to control them. But while allowing children some AI freedom, you must always remain vigilant about their online safety and healthy growth,” comments Andrey Sidenko, Cyber Literacy Projects Lead at Kaspersky.

Firm Reviews the Evolution of Phishing Threats in 2025

A new Kaspersky review reveals how cybercriminals revived and refined phishing techniques to target individuals and businesses in 2025, including calendar-based attacks, voice message deceptions and sophisticated multi-factor authentication (MFA) bypass schemes.

The findings emphasise the critical need for user vigilance, employee training and advanced email protection solutions to counter these persistent threats moving forward.

Calendar-based phishing targets office workers

A tactic originally from the late 2010s, calendar-based phishing, has reemerged with a focus on B2B environments. Attackers send emails with calendar event invitations, often containing no body text, hiding malicious links in the event description.

When opened, the event auto-adds to the user’s calendar, with reminders urging them to click links leading to fake login pages, such as those mimicking Microsoft.

Previously aimed at Google Calendar users in mass campaigns, this method now targets office employees. Organisations should conduct regular phishing awareness training, such as simulated attack workshops, to teach employees to verify unexpected calendar invites.

Voice message phishing with CAPTCHA evasion

Phishers are deploying minimalist emails posing as voice message notifications, containing sparse text and a link to a basic landing page. Clicking the link triggers a chain of CAPTCHA verifications to bypass security bots, ultimately directing users to a fraudulent Google login page that validates email addresses and captures credentials.

This multi-layered deception highlights the need for employee training programmes, such as interactive modules on recognising suspicious links and advanced email server protection solutions like Kaspersky SecureMail, which detect and block such covert tactics.

MFA bypass via fake cloud service logins

These sophisticated phishing campaigns are targeting multi-factor authentication (MFA) by mimicking services like pCloud (a cloud storage provider that offers encrypted file storage, sharing and backup services).

These emails, disguised as neutral support follow-ups, lead to fake login pages on lookalike domains (e.g., pcloud.online). The pages interact with the real pCloud service via API, validating emails and prompting for OTP codes and passwords, granting attackers account access upon successful login.

To counter this, organisations should implement mandatory cybersecurity training and deploy email security solutions like Kaspersky Security for Mail Servers, which flags fraudulent domains and API-driven attacks.

“With phishing schemes growing more deceptive, Kaspersky urges users to treat unusual email attachments, like password-protected PDFs or QR codes, with caution and verify website URLs before entering any credentials.

“Organisations should adopt comprehensive training programmes, which includes real-world simulations and best practices for spotting phishing attempts. Additionally, deploying robust email server protection solutions ensures real-time detection and blocking of advanced phishing tactics,” comments Roman Dedenok, Anti-Spam Expert at Kaspersky.