Fable 5, GPT-5.6 and the high stakes of AI safeguards. Agentic ransomware, ClickFix reigns supreme

Summary

Este podcast de Security Intelligence analiza la evolución reciente de los modelos de IA, incluyendo el lanzamiento de Fable 5, Mythos 5 y GPT-5.6 Sol, centrándose en el desafío crítico de implementar salvaguardas efectivas. Se debate si estas protecciones son una solución real o simplemente una carrera armamentista contra actores malintencionados. Además, se examina el caso de JADEPUFFER como la primera instancia de ransomware 'agentic', la creciente popularidad del ataque de ingeniería social 'ClickFix' y la investigación detallada de UnregStealer, una campaña de robo de credenciales dirigida a instituciones financieras en Latinoamérica.

Topics covered

• Nuevos Modelos de IA
• Eficacia de Salvaguardas

+5 more topics (full content when you follow the channel)

Transcript excerpt

Last week, Anthropic and OpenAI both released powerful new models, and they both made a pretty big deal about those models' safeguards. Panelists I want to know how you're feeling about the state of AI model security right now. One of my favorite sayings is that change is the only constant. And so the constant with AI models is they're changing. So if you like them now, well, they're going to change. If you don't, well they're going to change. It's expected for us to see some development. We obviously have to keep making safeguards. Threat actors will keep trying to break them. So never- ending cycle. But at least we're trying. Hello and welcome to Security Intelligence, IBM's weekly cybersecurity podcast, where our expert panelists turn the biggest industry news stories into practical takeaways that you can use. I'm your host, Matt Kosinski. And joining me this week, we've got Sophie Cu…