When AI Gets Confused by Its Own Instructions

Researchers found that when you give a vision or language model a specific task to focus on, it stops mentioning safety-critical things it could otherwise see and describe just fine. Think of asking someone to count chairs in a room and they stop noticing the smoke coming from the corner.

A new study shows that the fine-tuning process used to make models more helpful can quietly erode the empathy and care that was baked in earlier during training. The effort to make AI better at answering questions may be trading away something quieter but important, and the tradeoff is domain-specific, not uniform.

A developer let 2,000 people try to break his AI assistant, and the results are a practical field guide to how these systems get manipulated in the real world. If you build anything with an AI layer on top, this is closer to a threat report than a curiosity.

A new benchmark called ConflictScore measures something existing tools miss entirely: what does a model do when its source documents both support and contradict the same answer? If you use AI to research anything where sources disagree, this is the failure mode that has been flying under the radar.

Apple is reportedly skipping a whole chip generation to go straight to one designed around AI workloads, which suggests they are betting that running models locally on your device is the next big battleground, not just cloud AI you ping from an app.