The more targeted a cyber attack is, the more effective it can be. Andreas Wespi, Cyber Security Researcher at IBM, uses a specific example to explain the role played by AI procedures, "Hackers hide an encrypted malicious code in video conferencing software, for example. The code is only activated when a neural network trained on the basis of photos, which is also hidden in the video conferencing program, recognizes the target person." It is very difficult for normal anti-virus software to detect such malware, as it only becomes active immediately in the event of an attack. In order to identify such cleverly disguised malware, IBM therefore also uses AI procedures that specifically search for possible triggers for an attack.
In addition to artificial intelligence, quantum computers are also a particular challenge for data security. For example, the asymmetric public key algorithms that are often used for encryption nowadays, are not "quantum safe", i.e., they can generally be cracked by a quantum computer. However, it remains to be seen when such computers will achieve the necessary computing power in practice. Although new quantum-safe algorithms are already being developed, Wespi does not expect their standardization to be completed until 2023 or 2024. Since it is not yet clear which technology will ultimately prevail, IBM recommends cryptographic agility to companies. This refers to the fundamental ability to exchange all currently used cryptographic procedures for new standards at short notice as soon as they become available.
For Julian Meyrick, Vice President IBM Security Europe, all these technological developments are "a battlefield" on which the battle for data security is decided. At the same time, Cybersecurity is also a fast-growing business area for IBM, which currently employs around 8500 people.