AI and Cybersecurity 2026: EFF-Austin/Austin Science Network

RSVP at https://luma.com/whwsbusr Summer 2026 is buzzing with developments and commentary about AI risks and cybersecurity, especially the June 2026 US Government export control directive that suspends all access to Anthropic's Fable 5 and Mythos 5 models by any non US citizen or a foreign national, including Anthropic employees (this was followed by a partial-reverse order suspending access to Anthropic top models late in June). The Fable 5 model has guardrails as it "includes safety classifiers that can decline requests" while Claude Mythos 5 "does not include these classifiers." Meanwhile, open-source AI models such as from OpenRouter and DeepSeek, seek to gain adherence in an increasingly expensive and regulated AI world. In addition, the spectacular sums spent on capex (capital expenditures) by organizations on AI stimulates markets even as it both tempts worries investors. Speculation is swirling about these developments relative to hackers and cyber-risk, humans losing control of AI, geopolitics and the possibly massive IPO's coming for AI companies, making sober discernment…