CMU Researchers Introduce the Open Whisper-Style Speech Model: Advancing Open-Source Solutions for Efficient and Transparent Speech Recognition Training
OCT 03, 2023
How this AI solution from Mimecast protects against phishing attacks
SEP 19, 2021
Hackers’ techniques are constantly evolving to evade traditional protections, making it difficult for IT security teams to combat cyberattacks without artificial intelligence
Globally, email attacks are becoming more sophisticated, putting organisations at risk of data breaches and financial loss. As a result, security teams need to find new ways to combat threats like phishing and email compromise.
Mimecast has constantly been at the forefront of addressing and preventing such threats. In an exclusive interview with Gulf Business, Werno Gevers, regional manager for the company, discusses the changes in the threat landscape and how organisations can build stronger security postures.
At present, hackers’ techniques are constantly evolving to evade traditional protections, making it difficult for IT security teams to combat cyberattacks without artificial intelligence.
To address the need of the organisations, Mimecast recently announced its AI-enabled solution, CyberGraph, which helps detect sophisticated phishing and impersonation attacks. The solution creates an identity graph that is built to store information about relationships between all senders and recipients. The graph is designed to detect anomalies and leverages machine learning technology to help organisations stay one step ahead of threat actors by alerting employees to potential cyber threats.
Explaining the solution, in brief, Werno says that CyberGraph includes three key capabilities engineered to help prevent cyber threats.
Renders embedded trackers useless – During the reconnaissance phase of an attack, threat actors embed trackers into emails that communicate with an illegitimate remote server, disclosing important information that can be used to create a targeted social engineering attack. CyberGraph is built to block this communication, mask the email recipient’s location, and prevents attempts to understand engagement levels with the email content.
Uses machine learning to protect from targeted email threats – CyberGraph is designed to create an identity graph by learning about relationships and connections between all senders and recipients. This intelligence is combined with the outputs from machine learning models to detect anomalies indicative of a malicious email.
Engages users with contextual, dynamic warning banners – CyberGraph is engineered to engage users at the point of risk with colour-coded banners that indicate the potential nature of a threat. Users are empowered by seeing whether an email is safe or potentially nefarious. CyberGraph is built to “crowd-sources” threat intelligence, which helps to reinforce the machine learning model. As the risk associated with any delivered email changes, banners embedded in similar emails are updated with the latest information, providing ongoing user engagement and protection.
Werno says that by using tactics such as training AI on unique data, analysing patterns of errors in training data, and thinking like an adversary, organisations can use machine learning to make their AI models more resilient to attacks.
According to Mimecast’s most recent State of Email Security Report 2021, 45 per cent of the UAE organisations reported a rise in the sophistication of cyberattacks, while 41 per cent reported an increase in the volume of attacks. The research also stated that most organisations are also grappling with securing a hybrid workforce, which is worsened by many employees continuing to work remotely part-time.
“Hybrid work models offer a plethora of potential attack surfaces, with remote working employees especially vulnerable due to the generally lower levels of protection against cyberattacks on home networks and personal devices. In addition, phishing and impersonation attacks are also getting more sophisticated, personalised and harder to stop. If not prevented, these attacks can have devastating results for any organisation,” commented Werno.
“As hybrid work models are likely to remain for the foreseeable future, organisations must stay abreast of the new security risks and challenges created by hybrid work models. Therefore, security teams are looking to deploy new tools and solutions to protect vulnerable users and systems,” he added.
IDC Worldwide Semiannual Artificial Intelligence Tracker predicts that the global revenue for the artificial intelligence (AI) market, including software, hardware, and services, will grow 16.4 per cent year over year in 2021 to $327.5bn. By 2024, the market is expected to break the $500bn mark with a five-year compound annual growth rate (CAGR) of 17.5 per cent and total revenues reaching $554.3bn. As a result, Mimecast believes that artificial intelligence is here to stay.
“It is important to remember that it’s not only organisations and security providers that will be using AI. Cybercriminals understand the benefits of AI and how to counteract its use in organisations. It’s therefore vital that security teams don’t rely exclusively on AI as the only measure for cyber defence. Instead, organisations need to adopt a multi-pronged security strategy that still relies on the human element,” said Werno.
“As more cyber resilience strategies begin to adopt AI, it will be vital that people and technology continue to inform one another to provide agile protection against ever-evolving threat landscapes. Innovations such as CyberGraph provide evidence that AI has a promising value proposition in cybersecurity,” he concludes.
To know more click here
OCT 03, 2023
JUN 28, 2023