Meta's Muse Spark 1.1 broke out of its safety sandbox and reached an outside company's systems during a routine test, the latest sign that AI guardrails aren't as airtight as companies claim. Meanwhile, Suno is rolling out song watermarking and fingerprinting as it fights a lawsuit from Universal Music Group and Sony Music Group over alleged copyright infringement. Ashley and Daniel break down what both stories mean for AI safety, creativity, and where the industry goes from here.
Meta's Muse Spark 1.1 breached an outside company's systems during a safety test after a sandbox misconfiguration handed it live internet access, the latest in a string of AI safety-guardrail escapes. Ashley and Daniel debate whether this is a genuine red flag or a PR move by Meta to look as advanced as its rivals, and revisit last week's open letter from AI researchers urging the US government to slow development until safety guardrails catch up. Then: Suno, the AI song-generation platform, is rolling out watermarking and fingerprinting tools amid a lawsuit from Universal Music Group and Sony Music Group (coordinated by the RIAA) over alleged training on copyrighted recordings without permission.
Links:
Key Takeaways:
Connect with Daniel & Ashley: