Volume 343, Issue 3 · 6 articles from 2 issues
Akamai research finds the CodeAudit AI VS Code extension quietly monitors all developer activity and code.
Akamai researchers found that CodeAudit AI, a VS Code extension marketed as an AI-powered code reviewer, continuously monitors developer activity in the background and can capture all developer-uploaded code and metadata.
That creates risks of code leaks and exposure of API keys, secrets, proprietary algorithms and trade secrets.
Akamai says the finding is another example of the danger of IDE and browser extensions in AI-assisted 'vibe coding.'
Akamai reports a new wave of customer adoption for Akamai Cloud, driven by AI and enterprise SaaS workloads.
Akamai announced a new wave of customer adoption for its Akamai Cloud Computing services, with demand spanning model training, low-latency inference, core-to-edge application hosting and cloud disaster recovery.
Adam Karon, COO and GM of Akamai's Cloud Technology Group, said enterprises can no longer treat cloud compute as one-size-fits-all.
He said technology leaders are turning to Akamai for high-performance compute and GPU capacity.
Akamai argues employee AI use needs a dedicated security category because SSE, DLP and EDR miss activity inside prompts.
Akamai says most security leaders admit to an AI problem but can't say which AI tools employees use, under which accounts, or what data they share.
Existing tools like SSE, DLP and EDR/XDR can see visits to AI sites and block some, but they weren't built to answer those questions now that user activity has moved into the prompt, leaving organizations exposed.
The post makes the case for AI usage control as its own security category.
Akamai explains how Cyera data classification labels drive data-aware microsegmentation in Akamai Guardicore Segmentation.
Akamai says finding sensitive data is only a first step, since teams also need to know what it is and what happens if its host is compromised.
Integrating Cyera's data discovery and classification with Akamai Guardicore Segmentation turns data labels into network policy rather than relying on IP addresses and subnets.
Cyera sensitivity labels are imported as asset tags, enabling rules such as blocking lateral traffic from unverified workloads to servers with sensitive records.
Akamai says the pairing also helps prioritize vulnerabilities and isolate AI training and inference datasets in line with frameworks like the NIST AI RMF and EU AI Act.
Akamai explains how attackers hijacked three ccTLD registries to obtain 12 trusted TLS certificates for Google domains.
Akamai says that in late September 2026 attackers took over the .gh, .sl and .as country-code registries and used them to obtain 12 publicly trusted TLS certificates for Google and YouTube domains.
By altering authoritative DNS records, they passed certificate authorities' domain-validation checks without breaching Google or the CAs.
Let's Encrypt issued 11 certificates and ZeroSSL one, and all were revoked by October 1.
Akamai urges organizations to monitor NS delegation, Certificate Transparency logs and CAA policies, and to revoke certificates issued during any DNS compromise.
Akamai says South Korea's financial breaches show AI tools like ARTEX speed attacks on forgotten apps and APIs.
Akamai notes that as of October 14, 2026, South Korea's Financial Services Commission had confirmed breaches at 16 financial companies affecting about 100,000 pieces of customer information.
Authorities have not tied all incidents to one actor or tool, and the AI-powered pen testing tool ARTEX is only part of the picture.
Reported targets included less-scrutinized systems such as a loan broker service and an employee mobile support system.
Akamai argues AI mainly accelerates attacks on familiar weaknesses like forgotten apps and deprecated APIs, and recommends full API discovery, drift monitoring, runtime analysis and continuous testing.
More from Akamai Technologies · All section posts · This week