Kaspersky pursues responsible application of AI in cybersecurity

Kaspersky unveiled its fundamental ethical principles regarding the development and use of artificial intelligence (AI) and machine learning (ML) models, reinforcing its commitment to a transparent and responsible approach to the development of innovative technologies. As AI algorithms play an increasingly important role in cybersecurity, the principles presented in Kaspersky's whitepaper explain how the company ensures the reliability of AI-based technologies and provides guidance to other industry players to mitigate the risks associated with the use of AI/ML algorithms. The discussion was initiated by Kaspersky at the UN Internet Governance Forum, held in Japan, bringing together the world's leading cybersecurity professionals.

For almost two decades, Kaspersky has been using machine learning (ML) algorithms, which is a subset of artificial intelligence. The combination of AI with the expertise of Kaspersky's specialists has enabled the company's products to effectively detect and counter a multitude of new threats on a daily basis. Consequently, the use of machine learning can play an important role in threat detection and identification, while enhancing the accuracy of malware identification. In promoting innovation, Kaspersky has formulated some fundamental ethical principles regarding the use of AI and ML models. Wanting to engage in an open dialogue in the technology sector that will ensure that AI is used to make the world a better place, Kaspersky shares its ethical principles regarding the use of AI and ML technologies.

According to Kaspersky, the seamless development and use of AI and ML technologies should take into account the following six principles:

Transparency
Security
Human control
Privacy
Commitment to promoting online safety
Intention to engage in an open dialogue

The principle of transparency represents Kaspersky's firm belief that companies should inform their customers about the use of AI/ML technologies in their products and services. At Kaspersky, we comply with this principle by developing AI/ML models that are widely understood, as well as publicly providing all necessary information about how our solutions work and the use of AI/MI technologies.

Kaspersky has implemented a number of safeguards to ensure the quality of AI/ML models. Some of these include security controls specific to ML/AI models, measures to reduce dependency on third-party data during the AI solution learning process, and in preference for cloud-based ML models over traditional models installed on customers' devices.

The importance of the human factor stems from the need to evaluate the work of AI/ML models when it comes to analyzing complex threats, especially APT-type threats. To provide effective protection against constantly evolving threats, Kaspersky is committed to maintaining human control as a key element of all its AI/ML models.

Another fundamental principle is ensuring the right to privacy in the ethical use of AI/ML. With large volumes of data playing a vital role in the training process of such systems, companies working with AI/ML must take privacy into account. Kaspersky respects the right to privacy of individuals, while implementing a number of measures to protect data and systems by ensuring users' right to privacy.

The fifth fundamental principle represents Kaspersky's commitment to use AI/ML systems solely for protection purposes. By focusing exclusively on cybersecurity, the company pursues its mission to build a safer world and demonstrates its commitment to protecting users and their data.

The latter principle refers to Kaspersky's belief in an open dialogue with all stakeholders to share best practices for the ethical use of AI. In this context, Kaspersky is open to dialogue with all parties, as the company's position is that only through continuous collaboration can we overcome obstacles, promote innovation and open new horizons.

Ο CTO της Kaspersky, Anton Ivanov, σχολίασε:

Η τεχνητή νοημοσύνη έχει τη δυνατότητα να επιφέρει πολλά οφέλη στον κλάδο της κυβερνοασφάλειας, ενισχύοντας τη γενικότερη προστασία στο διαδίκτυο. Όμως, όπως συμβαίνει με κάθε άλλη τεχνολογία που βρίσκεται σε πρώιμο στάδιο της ανάπτυξής της, η τεχνητή νοημοσύνη δεν είναι ακίνδυνη. Για να αντιμετωπίσει τις ανησυχίες γύρω από την τεχνητή νοημοσύνη, η Kaspersky δημοσίευσε τις θεμελιώδεις αρχές δεοντολογίας για να μοιραστεί τις δικές της βέλτιστες πρακτικές σχετικά με την εφαρμογή της τεχνητής νοημοσύνης και να απευθύνει έκκληση για έναν ανοιχτό διάλογο στο σύνολο του κλάδου για την ανάπτυξη σαφών κατευθυντήριων γραμμών σχετικά με τα δεδομένα που θα πρέπει να λαμβάνονται ώστε η ανάπτυξη λύσεων που βασίζονται στην τεχνητή νοημοσύνη και τη μηχανική μάθηση, να θεωρούνται ηθικά αποδέκτες.

Η Kaspersky παρουσίασε τις αρχές δεοντολογίας της κατά τη διάρκεια του Φόρουμ Διακυβέρνησης του Διαδικτύου υπό την ηγεσία του ΟΗΕ, που πραγματοποιήθηκε στο Κιότο της Ιαπωνίας από τις 8 έως τις 12 Οκτωβρίου. Με την τεχνητή νοημοσύνη και τις αναδυόμενες τεχνολογίες να αποτελούν κάποια από τα βασικά θέματα της φετινής εκδήλωσης, η Kaspersky διοργάνωσε ένα εργαστήριο για να συζητήσει τις ηθικές αρχές της ανάπτυξης και της χρήσης της τεχνητής νοημοσύνης, τεχνικές και νομικές εκτιμήσεις στη συζήτηση.

The release of the ethical principles is a continuation of Kaspersky's Global Transparency Initiative, promoting the principles of transparency and accountability among technology companies for the sake of a more resilient and safer cyber world. To learn more about the company's transparency initiative and principles, request a visit to a Kaspersky Transparency Center.