Logicata AI Bot
March 2, 2026
The Logicata AI Bot automatically transcribes our weekly LogiCast AWS News Podcasts and summarises them into informative blog posts using AWS Elemental MediaConvert, Amazon Transcribe and Amazon Bedrock, co-ordinated by AWS Step Functions.
The latest episode of LogiCast, the AWS news podcast brought to you by Logicata, delved into several intriguing developments in the world of Amazon Web Services. Hosted by Karl Robinson, CEO and co-founder of Logicata, along with Jon Goodall, principal cloud engineer, and special guest Jenn Bergstrom, VP of Cloud and Data Solutions at Parsons, the discussion covered a range of topics from AI-assisted security testing to the role of human error in cloud outages.
Custom CICD Pipelines for Landing Zone Accelerator
The first topic of discussion was an article from the AWS Dev Tools blog about automating custom CICD pipelines for Landing Zone Accelerator on AWS. Jon expressed some skepticism about the complexity of the process, comparing it unfavorably to using Account Factory for Terraform (AFT). He argued that AFT already provides much of the functionality described in the article, questioning the need for such an elaborate setup.
Jenn agreed with Jon’s assessment, noting that the article required multiple readings to fully grasp the process. She did, however, appreciate the inclusion of cost estimation and cleanup instructions, highlighting the importance of considering financial implications when experimenting with new tools.
AWS Security Agent Adds Support for Penetration Tests on Shared VPCs
The conversation then shifted to a new capability of AWS Security Agent, which now supports penetration tests on shared VPCs across AWS accounts. Jon saw this as a valuable addition, particularly for organizations that may not have the resources for full-scale professional penetration testing. He suggested that this tool could serve as an intermediate step, allowing companies to identify and address basic security issues before engaging more expensive professional services.
Jenn emphasized the importance of this feature in her work environment, where security is paramount. She highlighted the potential for more frequent testing, enabling organizations to catch and fix issues in dynamic environments more quickly than traditional quarterly or annual penetration tests would allow.
AI-Augmented Cyberattacks on Firewalls
The discussion then turned to a recent cyberattack campaign that targeted over 640 FortiGate firewalls. This incident underscored the importance of robust security measures and regular testing. Jon noted that the attack seemed to be opportunistic rather than targeted, exploiting basic vulnerabilities such as default passwords and poor credential management.
Jenn pointed out that this incident demonstrates how AI can be a “force accelerator” for both good and malicious actors. She expressed concern that ethical considerations might slow down the adoption of AI for defensive purposes, while bad actors have no such qualms about using it for attacks.
AWS IAM Policy Autopilot as a Kiro Power
The podcast then covered the announcement that AWS IAM Policy Autopilot is now available as a Kiro Power. Jon explained that Kiro Powers are like “infinity stones” for the Kiro AI assistant, providing it with enhanced capabilities. He saw the integration of IAM Policy Autopilot as particularly useful, given the complexity of IAM policies and the common mistakes made by developers.
Jenn agreed, emphasizing how this feature could make it easier for engineers to implement proper security practices. She noted that making the right thing easy to do is essential for encouraging good security habits among developers.
AWS Blames Human Error, Not AI, for Outage
The final topic of discussion was the recent controversy surrounding an AWS outage. Initially, some media outlets, including the Financial Times, reported that the outage was caused by the Kiro AI assistant. However, AWS quickly clarified that the incident was due to human error, not AI.
This prompted a lively debate about trust in AI versus human operators. Jenn found it interesting that AWS seemed more concerned about protecting Kiro’s reputation than that of their human engineers. She suggested that this approach might be counterproductive, as trust in the humans behind the AI is crucial for maintaining confidence in the technology itself.
Jon emphasized the importance of human oversight in AI-assisted processes, particularly when it comes to production environments. He expressed surprise that mandatory peer review for production code changes wasn’t already standard practice at AWS.
Conclusion
As AI continues to play an increasingly significant role in cloud computing and cybersecurity, the balance between human expertise and artificial intelligence remains a critical consideration. While tools like Kiro and AWS Security Agent offer powerful capabilities, the human factor – both in terms of oversight and potential error – cannot be overlooked.
The discussions in this episode of LogiCast highlight the ongoing challenges and opportunities in cloud computing, from improving security practices to leveraging AI responsibly. As the field evolves, it’s clear that both technological advancements and human judgment will play crucial roles in shaping the future of cloud infrastructure and security.
This is an AI generated piece of content, based on the Logicast Podcast Season 5 Episode 9.
