AI-Driven Phishing Detection Systems

Leveraging the power of artificial intelligence, women engineers are at the forefront of developing systems that can learn from vast amounts of data to recognize the characteristics of phishing emails and websites. These systems continually adapt to the evolving tactics of cybercriminals, providing real-time alerts to users about potential threats.

Leveraging the power of artificial intelligence, women engineers are at the forefront of developing systems that can learn from vast amounts of data to recognize the characteristics of phishing emails and websites. These systems continually adapt to the evolving tactics of cybercriminals, providing real-time alerts to users about potential threats.

Empowered by Artificial Intelligence and the women in tech community.
Like this article?
Sun, 12/08/2024 - 23:58

Recent research on phishing URL detection highlights the use of deep learning models, particularly Convolutional Neural Networks (CNNs), often in combination with other techniques like Long Short-Term Memory (LSTM) networks or Hybrid models, to achieve detection accuracies exceeding 90%.

However, key challenges remain, including high false-positive rates, vulnerability to adversarial attacks, and limited real-time assessment of training data. Many models are trained on outdated data, missing critical features such as HTML, n-grams, and image data. The literature also points to a gap in incorporating multimodal data (e.g., website screenshots, HTML content) to address visual deception techniques used in phishing attacks.

Future advancements in phishing URL detection will focus on improving model robustness through adversarial training, incorporating multimodal data, real-time deployment, and enhancing transparency with explainable AI (XAI). Collaboration between researchers, cybersecurity experts, and industry stakeholders is essential to addressing these challenges and improving detection systems to counter evolving phishing threats.

Link to the full paper: https://www.tandfonline.com/doi/full/10.1080/23742917.2024.2378552

...Read more
0 reactions
.
Contribute to three or more articles across any domain to qualify for the Contributor badge. Please check back tomorrow for updates on your progress.

Interested in sharing your knowledge ?

Learn more about how to contribute.