Banned is not the same as safe
In short
AAnomaly
In October 2026, Panasonic reported that one employee uploaded personal data of about 334,000 people into an outside AI service. That service was banned 1. There was even an in-house AI tool. Two months earlier, the same part of Panasonic said that tool could not cover every need 2.
RReframe
The headlines are about one person breaking a rule. I am more curious about the rule. For someone with a hundred open tasks, the faster tool wins. A ban tells people where not to go. It does not give them a road.
CCreate: questions for your team
- Do our rules fit how people really work?
- Does the tool we gave them actually help?
Rethink the premise.
Go deeper The full piece, with all sources
AAnomaly
On 23 October 2026 Panasonic reported that one project member had uploaded personal data of about 334,000 people, staff, former staff and their families, into an outside generative AI service. That service was banned. There was no hacker 1. Two months earlier, the same part of Panasonic had said in a case study that its own AI tool, used by around 80,000 people, could not cover every need on its own 2. Why the employee did it has not been made public, and I won't guess.
RReframe
The headlines are about one person breaking a rule. I am more curious about the rule. For a board, the risk is obvious. For someone with a hundred open tasks, the faster tool wins. And an AI does not feel like a stranger. It feels like a tool. A ban tells people where not to go. It does not give them a road.
CCreate: questions for your team
- Do our rules fit how people really work?
- Do people know enough to see the risk?
- Does the tool we gave them actually help?
Rethink the premise.
Sources
- Statement on personal data uploaded to an external generative AI serviceStrong
- Case study PX-AI adoptionMedium · vendor source
- Datalek bij PanasonicStrong
Want to work on this with your team? See our products →