Adobe uncovered Project Voco a week ago. The product makes it conceivable to take a sound recording and quickly modify it to incorporate words and expressions the first speaker never articulated, in what sounds like their voice.
One master cautioned that the tech could promote undermine confide in news coverage.
Another said it could represent a security danger.
Nonetheless, the US programming firm says it is making a move to address such dangers.
At a live demo in San Diego on Thursday, Adobe took a digitized recording of a man saying "and I kissed my canines and my better half" and transformed it to say "and I kissed Jordan three times".
The alter took seconds and just included the administrator overtyping a transcript of the discourse and after that squeezing a catch to make the combined voice track.
"We have as of now upset photograph altering. Presently it's the ideal opportunity for us to do the sound stuff," said Adobe's Zeyu Jin, to the commendation of his group of onlookers.
He added that to make the procedure conceivable, the product should have been furnished with around 20 minutes-worth of a man's discourse.
Dr Eddy Borges Rey - an instructor in media and innovation at the University of Stirling - was frightened by the advancement.
"It appears that Adobe's developers were cleared alongside the fervor of making something as creative as a voice controller, and overlooked the moral issues raised by its potential abuse," he told the BBC.
"Coincidentally, in its journey to make programming to control computerized media, Adobe has [already] radically changed the way we connect with evidential material, for example, photos.
"This makes it hard for legal advisors, writers, and different experts who utilize advanced media as confirmation.
"Similarly that Adobe's Photoshop has confronted lawful reaction after the proceeded with abuse of the application by publicists, Voco, if discharged industrially, will take after its antecedent with comparable results."
The dangers reach out past individuals being tricked into supposing others said something they didn't.
Banks and different organizations have begun utilizing voiceprint checks to confirm clients are who they say they are the point at which they telephone in.
One cybersecurity specialist said the organizations included had since quite a while ago foreseen something like Adobe's development.
"The innovation is new yet its basic standards have been comprehended for quite a while," said Dr Steven Murdoch from University College London.
"Biometric organizations say their items would not be deceived by this, in light of the fact that the things they are searching for are not similar things that people search for when distinguishing individuals.
"In any case, the best way to discover is to test them, and it will be some time before we know the reply."
Google's DeepMind division flaunted an opponent voice-imitating framework called WaveNet in September.
In any case, at the time, it proposed that the errand required an excess of handling energy to discover its way into a purchaser item soon.
As far as concerns its, Adobe has discussed its clients utilizing Voco to settle podcast and book recording recordings without having to rebook moderators or voiceover craftsmen.
In any case, a representative focused on this did not mean its discharge was approaching.
"[It] might be discharged as an item or item include," she told the BBC.
"No ship date has been declared."
Meanwhile, Adobe said it was exploring approaches to distinguish utilization of its product.
"Consider watermarking location," Mr Jin said at the demo, alluding to a strategy used to shroud identifiers in pictures and other media.
Adobe Voco 'Photoshop-for-voice' causes concern