AI Cybersecurity

Today, intellectual property such as patent details, technological prototypes, and trade secrets are mostly stored in the digital network.
The broad categories of intellectual property are also stored in the electronic network.
They include technology blueprints, proprietary manufacturing processes, test results, business plans, pricing documents, partnership agreements, emails and contact lists from victim organization’s leadership.
Almost all the advantages are on the side of the hacker; hence, the current situation is not sustainable. Moreover, entirely defensive measures are likely to continue to become increasingly expensive and decreasingly effective, while being unlikely to change the cost-benefit calculus of targeted hackers away from attacking corporate networks

In this context, the security concept of threat-based deterrence and its effective tool are designed to dissuade the attackers from taking unwanted actions in the first place because they believe that they will incur unacceptably high costs and/or receive insufficient benefits from taking that action.
In other words, it reverses the time, opportunity, and resource advantage of the targeted attacker by reducing his incentives and raising his costs without raising costs for the defender.

We’re generating fakes by using “word embedding” which is using large corpuses of words.
Codes and tools for learning word embedding are already available to us but we need to train it by using documents.
The larger the corpus, the better the product.