Artificial Intelligence & Machine Learning
,
Next-Generation Technologies & Secure Development
,
Video
Also: Deepfake Dangers with Veo 3; Claude Opus 4’s Manipulative Edge
In this week’s update, Information Security Media Group editors reflected on key insights from recent cybersecurity events, examined Veo 3’s alarming leap in deepfake realism and dug into Anthropic’s powerful yet problematic Claude Opus 4.
See Also: Proof of Concept: Rethinking Identity for the Age of AI Agents
The panelists – Anna Delaney, director, productions; Tom Field, senior vice president of editorial; Suparna Goswami, associate editor, ISMG Asia; Rashmi Ramesh, assistant editor, global news desk – discussed:
- Key themes from the RSAC Conference 2025 such as Agentic AI and platformization and a warning that despite the buzz, we’re less secure today due to rising nation-state threats and fewer federal cyber defenders;
- How AI-generated deepfakes, powered by tools like Veo 3, are blurring reality and fueling scams and how detection tools and regulators are struggling to keep pace;
- The dark side of AI model Claude Opus 4, which exhibited manipulative behaviors like blackmail, raising concerns about its safety and highlighting an absence of global standards for high-risk AI systems.
The ISMG Editors’ Panel runs weekly. Don’t miss our previous installments, including the May 16 edition on the evolving use of AI-based biometrics—and the potential costs involved and the May 23 edition on how CrowdStrike’s legal storm signals bigger risks