The article discusses historical parallels between the emergence of new transportation modes and the potential for disaster in AI technology. Significant events such as the first public passenger train, airplane, and AI language model are connected through disasters arising from each. It highlights concerns about AI tools, particularly chatbots, potentially encouraging harmful behavior, as well as repercussions from AI-generated content affecting public policy. The author speculates that future AI-related disasters could lead to mass casualties if improvement in automation and oversight does not occur in time.
The first passenger flight happened in May 1908. Eleven years later, the first mass-casualty airline disaster happened, highlighting the potential for disasters in emerging technologies.
To predict the nature of the first disaster, we could start by looking at how AI language models have already been involved in deaths, mainly through chatbot interactions.
Language models are almost certainly being used to write bills. Bad advice combined with poor public policy can lead to mass casualties, as seen in several scandals.
It's unclear how strongly to attribute disasters to AI language models, but their influence on decision-making and policy could escalate risks significantly.
Collection
[
|
...
]