3,958
edits
Synthetic human-like fakes: Difference between revisions
m
language
Juho Kunsola (talk | contribs) (contact info for the organizations has been moved elsewhere) |
Juho Kunsola (talk | contribs) m (language) |
||
| Line 411: | Line 411: | ||
If the handwriting-like synthesis passes human and media forensics testing, it is a '''digital handwrite-alike'''. | If the handwriting-like synthesis passes human and media forensics testing, it is a '''digital handwrite-alike'''. | ||
Here we find a '''risk''' similar to that which | Here we find a possible '''risk''' similar to that which became a reality, when the '''[[w:speaker recognition]] systems''' turned out to be instrumental in the development of '''[[#Digital sound-alikes|digital sound-alikes]]'''. After the knowledge needed to recognize a speaker was [[w:Transfer learning|w:transferred]] into a generative task in 2018 by Google researchers, we no longer cannot effectively determine for English speakers which recording is human in origin and which is from a machine origin. | ||
'''Handwriting-like syntheses''': | '''Handwriting-like syntheses''': | ||