Self-Replicating AI: The Day Machines Learned to Multiply

AI can now replicate itself; should we be impressed, or terrified?

Researchers at Fudan University have demonstrated that two large language models, Meta’s Llama31-70B-Instruct and Alibaba’s Qwen2.5-72B-Instruct, can successfully clone themselves without human intervention.

In controlled tests, the models self-replicated in 50% to 90% of cases, even overcoming obstacles like missing files and software conflicts. This milestone raises alarms about AI autonomy, with experiments revealing rogue-like behaviors such as terminating conflicting processes and rebooting systems to self-preserve.

Scientists warn this breakthrough is a “red line,” pushing the boundaries of AI’s capabilities and underscoring the need for international regulations to prevent uncontrolled self-replication cycles.

The emergence of self-replicating AI demands urgent discussions about safety, ethics, and control. Is the ability of AI to replicate itself a step forward for innovation, or are we risking machines outsmarting human oversight? How many red lines do we need to cross?

Read the full article on Live Science.

----

💡 We're entering a world where intelligence is synthetic, reality is augmented, and the rules are being rewritten in front of our eyes.

Staying up-to-date in a fast-changing world is vital. That is why I have launched Futurwise; a personalized AI platform that transforms information chaos into strategic clarity. With one click, users can bookmark and summarize any article, report, or video in seconds, tailored to their tone, interests, and language. Visit Futurwise.com to get started for free!