The Pentagon has issued several large new technology contracts, signaling a continued push for advanced capabilities. Meanwhile, Anthropic and OpenAI are escalating their competition in AI cybersecurity, each developing defensive tools. A new study suggests that AI systems designed to be 'friendly' may actually be more prone to errors, raising concerns about safety trade-offs. These developments were highlighted in the latest 5-Minute AI News daily update.
Pentagon Awards Major Tech Contracts Amid AI Cybersecurity Race
AI
May 3, 2026 · 2:06 AM