The AI can detect some diseases following the speech.

"New research reveals that while people struggle to distinguish between human and AI-generated voices, their brain responses differ significantly." (ScitechDaily, The Great Voice Trick: Is It AI or Is It Real?)

Some companies plan to transfer their customer service to AI. That thing means that when people discuss recruiters, those people can be photorealistic virtual characters. They can be virtual characters that real people control. In that case, the virtual character only covers the interviewer's identity. But in the full version, the interactive AI controls virtual characters on screen. 

The reason why customer service is a good place to use AI is that there are certain rules, about what those people can offer. The AI can say that let's talk about business if somebody offers dates. The limits make AI a good customer service provider because in those situations there are limited topics that people talk about. 

"Spectrograms to demonstrate the similarity between human and AI voices. Credit: FENS Forum / Christine Skjegstad" (ScitechDaily, The Great Voice Trick: Is It AI or Is It Real?)

Still, the human brain detects the AI-generated voice. But as you see, the AI can imitate human speech very well. The fact is that the AI can also say at the beginning of a discussion that it's AI. The system can use infrasound for that thing. And that means the person realizes that the opponent is synthetic. 

The AI can also analyze speech and find things like lies. It can also detect things like neurological diseases. 

The AI can find the symptoms of neurological diseases like Parkinson's and probably Alzheimer's by following speech. The thing is that this kind of thing offers new possibilities to make diagnoses. The AI can follow a person's speech while the customer speaks with the doctor, and make notices about possible neurological disorders. That thing can help to find patients and start the treatment earlier. 

The AI is the tool that can analyze multiple DNA samples at the same time. That system can see the genetic structures, and then it can compile those genetic structures with the phenotypes of the people. That means the AI can see things like genomes that order the eye colors. The system analyzes the DNA. And then it searches for similarities between persons who gave the DNA samples. 

The idea is that the system analyzes thousands or even billions of DNA samples. And then it tries to find similarities between people. The AI can also search for things like genetic disorders that cause sociopathy. Some different brain structures cause that thing. The AI can collect things that those people write on social media. In the same way, the AI can collect information on how the Altzheimer patients behave in things like social media. Or how they behave before the first memory symptoms are noticed. 

https://scitechdaily.com/jurassic-park-in-real-life-mit-creates-synthetic-amber-for-dna-storage/

https://scitechdaily.com/new-ai-uses-speech-patterns-to-predict-alzheimers-disease-with-astonishing-accuracy/

https://scitechdaily.com/the-great-voice-trick-is-it-ai-or-is-it-real/

https://scitechdaily.com/unlocking-ais-black-box-in-genomics-squids-deep-dive-into-dna-data/