Consider just a few ways that AI can be used to create fraud through omission, latency, commission, oversampling, or bias. We already have AI created fake voices and fake images. This is just the beginning.
A. First of all, much of
the purported AI is little more than standard statistical analysis, with error
criteria unmentioned at best or concealed at worst.
B. Error that creep in to
AI results are forever. There seems no continuing validation. Very important in
medicine.
C. AI results and
conclusions may dominate human conclusions and decisions, deprecating the human factor. Already students are using AI to generate original assignments.
D. Generative AI can be
used to create new (i.e. fake) data and images, which in turn can be sold to
other data centers as original data. This will impact business decisions,
government regulations, and even scientific results - to mention just a few.
I expect in the next few years, the issues of fakes and fraud with AI will explode on the world stage. There soon may be senior ministers in governments, institutions, and military to monitor AI fraud.
Comments
Post a Comment
Please Comment.