Artificial intelligence adversary red team
An AI adversary red team configured to pentest email and/or network defenses implemented by a cyber threat defense system used to protect an organization and all its entities. AI model(s) trained with machine learning on contextual knowledge of the organization and configured to identify data points from the contextual knowledge including language-based data, email/network connectivity and behavior pattern data, and historic knowledgebase data. The trained AI models cooperate with an AI classifier in producing specific organization-based classifiers for the AI classifier. A phishing email generator generates automated phishing emails to pentest the defense systems, where the phishing email generator cooperates with the AI models to customize the automated phishing emails based on the identified data points of the organization and its entities. The customized phishing emails are then used to initiate one or more specific attacks on one or more specific users associated with the organization and its entities.
1 . An apparatus, comprising:
one or more trained Artificial Intelligence (AI) models configured to identify data points from contextual knowledge of an organization; and
a phishing email generator configured to generate one or more automated phishing emails to pentest an email defense system by at least (i) operating with the one or more trained AI models to customize the one or more automated phishing emails based on the identified data points of the organization and (ii) initiating, using the one or more customized automated phishing emails, at least a specific attack on one or more specific users associated with the organization, the phishing email generator comprises a paraphrasing engine configured to receive email data from the identified data points and separate the received email data into a plurality of segments of an email that include at least two or more of a subject line, a body content, and a signature line,
wherein the paraphrasing engine is further configured to rephrase the received email data so that one or more particular segments of a first phishing email from the one or more customized automated phishing emails is different from one or more particular segments of a second phishing email from the one or more customized automated phishing emails, and
wherein instructions implemented by the one or more AI models, the phishing email generator, and the paraphrasing engine, are configured to be stored in an executable format on one or more non-transitory computer readable medium and are configured to be executed by one or more processors.
2 . The apparatus of claim 1 , wherein the contextual knowledge includes language-based data, email and network connectivity and behavior pattern data, and historic knowledgebase data.
3 . The apparatus of claim 1 further comprising:
a payload module implemented as instructions stored in an executable format on the one or more non-transitory computer readable medium and executed by the one or more processors and configured to cooperate with the phishing email generator to generate at least one or more of a first payload and a second payload attached to the one or more customized automated phishing emails, wherein the first payload is configured as a non-executable payload and the second payload is configured as an executable payload.
4 . The apparatus of claim 3 further comprising:
a training module implemented as instructions stored in an executable format on the one or more non-transitory computer readable medium and executed by the one or more processors and configured to cooperate with the payload module to train the one or more specific users in the organization that activated the first payload attached to the one or more customized phishing emails.
5 . The apparatus of claim 4 further comprising:
a simulated cyber-attack module implemented as instructions stored in an executable format on the one or more non-transitory computer readable medium and executed by the one or more processors and configured to use the second payload attached to the one or more customized automated phishing emails to pentest a network defense system, wherein the simulated cyber-attack module is configured to cooperate with the one or more trained AI models to customize the one or more specific attacks in light of one or more specific attack scenarios in the network defense system.
6 . The apparatus of claim 1 further comprising:
an artificial intelligence (AI) adversary red team simulator implemented as instructions stored in an executable format on the one or more non-transitory computer readable medium and executed by the one or more processors and configured to coordinate the pentest of one or more defenses implemented by a cyber threat defense system, wherein the one or more defenses include at least the email defense system or a network defense system used to protect the organization.
7 . The apparatus of claim 6 being communicatively coupled to a cyber security appliance, wherein the cyber security appliance includes a profile manager module implemented as instructions stored in an executable format on the one or more non-transitory computer readable medium and executed by the one or more processors and configured to (i) communicate and cooperate with the AI adversary red team simulator, (ii) maintain a profile tag on each entity of the organization connecting to a network under analysis based on email and network connectivity and behavior pattern data, and (iii) supply the profile tag for entities connecting to the network.
8 . The apparatus of claim 6 further comprising:
a user interface implemented as instructions stored in an executable format on the one or more non-transitory computer readable medium and executed by the one or more processors and configured to cooperate with an orchestration module to provide one or more user input parameters specifically tailored to the organization and specified by a particular user in the organization,
wherein the orchestration module is implemented as instructions stored in an executable format on the one or more non-transitory computer readable medium and executed by the one or more processors and configured to reside on at least the AI adversary red team simulator, and
wherein the one or more user input parameters include (a) a first parameter configured to identify a predetermined attack to pentest the cyber threat defense system, (b) a second parameter configured to select a predetermined user and entity to be attacked with the identified predetermined attack, (c) a third parameter configured to establish a predetermined threshold to execute the identified predetermined attack on the selected predetermined user and the entity, and (d) a fourth parameter configured to restrict one or more predetermined users or one or more entities in the organization from being attacked, and
wherein the predetermined threshold is configured based on at least one or more of a predetermined time schedule allowed for that attack, a predetermined number of paths allowed for that attack, and a predetermined number of compromised users, devices, and entities prior to the pentest being allowed for that attack.
9 . A method for generating AI automated phishing emails to pentest a cyber threat defense system, comprising:
training one or more Artificial Intelligence (AI) models with machine learning on contextual knowledge of an organization, wherein the one or more trained AI models are configured to identify data points from the contextual knowledge of the organization; and
generating, by a phishing email generator, one or more automated phishing emails to pentest an email defense system by at least (i) operating with the one or more trained AI models to customize the one or more automated phishing emails based on the identified data points of the organization and (ii) initiating, using the one or more customized automated phishing emails, at least a specific attack on one or more specific users associated with the organization;
receiving email data from the identified data points by a paraphrasing engine of the phishing email generator and separating the received email data into a plurality of segments of an email that include at least two or more of a subject line, a body content, and a signature line; and
rephrasing the received email data so that one or more particular segments of a first phishing email from the one or more customized automated phishing emails is different from one or more particular segments of a second phishing email from the one or more customized automated phishing emails.
10 . The method of claim 9 , wherein the contextual knowledge includes language-based data, email and network connectivity and behavior pattern data, and historic knowledgebase data.
11 . The method of claim 9 further comprising:
generating, by a payload module configured to cooperate with the phishing email generator, at least one or more of a first payload and a second payload attached to the one or more customized automated phishing emails, wherein the first payload is configured as a non-executable payload and the second payload is configured as an executable payload.
12 . The method of claim 11 further comprising:
training the one or more specific users in the organization that activated the first payload attached to the one or more customized phishing emails.
13 . The method of claim 12 further comprising:
using the second payload attached to the one or more customized automated phishing emails to pentest a network defense system being different from the email defense system.
14 . The method of claim 9 , wherein prior to generating the one or more automated phishing emails, the method further comprising:
coordinating, by an AI adversary red team simulator, the pentest of one or more defenses implemented by a cyber threat defense system, wherein the one or more defenses include at least the email defense system or a network defense system used to protect the organization.
15 . The method of claim 14 further comprising:
establishing, by a profile manager module, communications with the AI adversary red team simulator;
maintaining a profile tag on each entity of the organization connecting to a network under analysis based on email and network connectivity and behavior pattern data; and
supplying the profile tag for entities connecting to the network.
16 . The method of claim 15 further comprising:
providing one or more user input parameters specifically tailored to the organization and specified by a particular user in the organization, wherein the one or more user input parameters include (a) a first parameter configured to identify a predetermined attack to pentest the cyber threat defense system, (b) a second parameter configured to select a predetermined user and entity to be attacked with the identified predetermined attack, (c) a third parameter configured to establish a predetermined threshold to execute the identified predetermined attack on the selected predetermined user and the entity, and (d) a fourth parameter configured to restrict one or more predetermined users or one or more entities in the organization from being attacked, and
wherein the predetermined threshold is configured based on at least one or more of a predetermined time schedule allowed for that attack, a predetermined number of paths allowed for that attack, and a predetermined number of compromised users, devices, and entities prior to the pentest being allowed for that attack.
17 . A non-transitory computer readable medium comprising: one or more computer readable codes operable, when executed by one or more processors, to instruct AI-based software to perform the method of claim 9 .
18 . An apparatus, comprising:
one or more processors; and
one or more non-transitory storage mediums including software configured for processing by the one or more processors content including
one or more Artificial Intelligence (AI) models configured to identify data points from contextual knowledge of an organization, and
a phishing email generator configured to generate one or more automated phishing emails to pentest an email defense system by at least (i) operating with the one or more trained AI models to customize the one or more automated phishing emails based on the identified data points of the organization and (ii) initiating, using the one or more customized automated phishing emails, at least a specific attack on one or more specific users associated with the organization, the phishing email generator comprises a paraphrasing engine configured to receive email data from the identified data points and separate the received email data into a plurality of segments of an email that include at least two or more of a subject line, a body content, and a signature line,
wherein the paraphrasing engine is further configured to rephrase the received email data so that one or more particular segments of a first phishing email from the one or more customized automated phishing emails is different from one or more particular segments of a second phishing email from the one or more customized automated phishing emails.