#the VAs in the case of AI did not agree to that and are all very vocal about not wanting their voices sampled

[venacoeurva]

I feel kind of insane that it was a very strong consensus that AI voices are unethical because the original VAs never consented to their voices being used in this way, are not compensated, and their voices are being used for things they never agreed/would never agree to say and it doesn't matter if these uses are commercial in nature or not, it feels very ghoulish, but one of the biggest trends rn is AI covers of songs and they're being pumped out like crazy and people are falling over themselves to defend people who don't have big time production budgets using them instead of hiring someone on Fiverr or imitating the voices themselves or voice clip mixing and sampling pre-existing lines like the olden days. ok. alright. Good to see some of you out there have the moral fortitude of a wet paper straw

[tyrianluda]

(Context: This is part of the Narrator essay I'm writing but I think this section was broad to taken outside of the essay. Take this as a little PSA)

(A Short Word about Boundaries and Parasocial Relationships: This is a Serious Section; FINISHED)

(Also a quick note: If you see a post or person that does something similar to what I'm writing about, please do not harass them. Especially when it's people having parasocial relationships. They just need to be careful to not have a unhealthy attachment over people since it hurts them mentally in the long-term. I chose to not name anyone or use specific pronouns for the sake of anonymity; not for any malicious attempt.)

For people in the fandom who are wondering, I have unfortunately seen people border on the line of sexualizing the Narrator’s VA, Kevan Brighting. More specifically, disregards his boundaries for how he wants his voice to be used and not used; especially in context to…not safe for work uses of his voice. I want to say this now before this escalates into people genuinely crossing this line and I feel like it’s a good reminder. We should care about making Kevan Brighting as comfortable as possible when engaging in the fandom, as he now knows what we say about him. As funny as his “sexiest man in Britain” tweet was, he knows. We should care because he does connect himself quite often in fan-made and Stanley Parable-inspired projects. He has always absorbed himself in the fanbase when he didn’t absolutely need to since the original game came out, and it’s best to have him stay comfortable doing this. Since his level of involvement in the fandom is not very common, to begin with, it’s not wise to give him a reason to stop. Also, more generally because it’s not okay at all to disregard anyone’s expressed boundaries and it shouldn’t be a normal thing to do. Actually, in this case, it’s really creepy and disrespectful and did I mention creepy?

For context, a Tumblr user by the name of @/shinakazami1 contacted Brighting to ask for his thoughts on his voice being used for AI. This was his response;

“...Regarding AI - I hate everything about it. The idea that someone could use my voice for something I don’t agree with is abhorrent. A voice artist must have artistic control over the work he/she does…” - Excerpt of his response

Tumblr Post; for Context: https://www.tumblr.com/vega-482/711812424798535680/kevan-brightings-opinion-on-voice-ai?source=share

You must be wondering about the connection between AI and the sexualization of the voice actor. See, the reason the user contacted Brighting, to begin with, was because another person on a Stanley Parable Discord server posted an NSFW audio of the Narrator. Using the AI of Brighting’s voice to do so. Pain. Agony, if you will.

There are a lot of ethical issues behind this action. Aside from the reason Brighting stated directly, this could also affect him financially; as he is a voice actor and general voiceover artist. AI audios of his voice would certainly impede his career. As an artist, this is a genuine thing he would probably worry about; and rightfully so. In the case of NSFW audio, this may also overstep a potential boundary he may have. Especially when it is NSFW without his permission. Even if he had no issue over NSFW audios particularly, he would have an issue about being in audio where he didn’t even get the chance to accept/deny. Even if you’re embarrassed, just ask him about his boundaries. Like, he should at least have a choice to say ‘no, fuck off’ and take requests on his own terms. Hopefully for obvious reasons. …but people kinda suck.

Unfortunately, it’s not just that one Discord user. Bro, I’m at my limit. I’m gonna have to talk about TikTok. God damnit. I said TikTok was its own beast about the Narrator for a reason. TikTok as well as Instagram, are mostly known in fandom spaces for containing edits of whatever character in whatever media. There’s a lot of work and effort that is put into making edits, and many are rewarded by people barking in the comment sections. As someone who has seen edits from different fandoms for about three to four years, I respect many editors and I understand why some would edit certain characters. Of course there are going to be edits of the Narrator; especially if you include Sad-ist’s Clock 0ut animations that are inspired by the Stanley Parable. It’s a hot old guy with a hot voice; I’m not exactly shocked by that in itself. By extension, of course some of those edits are suggestive or lean into NSFW territory. It’s understandable and it can be fine to make those edits if it doesn’t impede on real people’s boundaries or if the involved people openly say ‘It’s good, bro’. Now here’s where the problems start. 

At least a few TikTok edits have used AI of Brighting’s voice for parts that have NSFW. Whether it’s having the AI say NSFW things, having it be implied or even having it whimper, the main gist is that his voice is used for NSFW; despite the voice actor not directly consenting to a part of it. I’m so done with you people. Many potential problems would occur because of the decision of doing such. I already brought up how Brighting’s career of voicework could be affected financially by the use of AI but there’s another potential issue. It could probably also affect him socially and his involvement in future projects. The fact that these edits featuring AI don’t get that popular, is kind of a miracle. If those edits ever went outside of the fandom and into a more professional setting, it would be unnecessarily difficult for him to explain that he wasn’t even a part of those edits. I say ‘unnecessarily’ because you don’t even need AI to hear him in that context. 

You could’ve used a scene from one of his audiobooks. They work for their intended purpose and you don’t even have to buy the books. I understand not doing some things because of money problems but… NSFW audios aren’t a necessity and the audiobooks are free! If it’s because of convenience, then how is AI any more convenient? It’s not hard to find the audiobooks at all! Some of the audio files are on Tumblr for fuck’s sake. There’s no logical sense of why the decision of using AI is okay because it’s not okay. It is very disrespectful to disregard his boundaries and if I really wanted to go there, this borders on harassment. It’s not difficult to respect open boundaries that are openly stated with absolutely no ambiguity. Look, at least for the audiobooks, Brighting did them willingly. You did not need AI; you really didn’t. 

Some people even go as far as to say his tweet about him being the sexiest man in Britain was indicative that he’s fine with the sexualization of the Narrator and by extension, NSFW edits featuring AI. Well, that’s a fucking reach if I ever saw one. Even though the sexualization of the Narrator wasn’t even the issue stated, Brighting responded in an email stating how much he hated AI. The AI is the issue; not the Narrator simping. Kevan Brighting can think the sexualization of the Narrator is funny because the Narrator isn’t real. But Brighting is real. He’s a living, breathing, voice-acting guy with boundaries. You couldn’t be bothered to have the NSFW audios be from the audiobooks he agreed to be a part of and consented to do? If you could put enough effort into making a TikTok edit, you could put enough effort into not being a smooth brain. You don’t need a fictional character’s consent, but you do need consent from real-life people who are affected if you don’t respect their consent. If you can’t understand the difference between a fictional character and a real person to where it affects real people in reality, you shouldn’t be trusted to post anything of that nature.

It’s one thing to not know and later to apologize for your mistake. But, it’s another thing to hear his response that he doesn’t like AI of his voice period…like at all and it’s just…one ear and out the other. The people who make these edits with AI did not understand why he said this and should have apologized or deleted them after being disclosed of his boundaries. It won’t kill you to have basic courtesy. I assume most people don’t know of his thoughts on AI before making the edits because if they knew already, it would make the situation so much worse. Still, it’s worth an apology. (See Also: Hanlon's razor)

(Parasocial Relationship Part)

AI is not the only issue since TikTok also blurs the line when talking about the Narrator and Kevan Brighting in some unrelated cases. Some people forget that Brighting is a real guy and they think of him as like a fictional character.  

I saw a TikTok post that wondered why Brighting was wearing women’s frames for his glasses in one image and half the comments were like ‘bully him’. First of all, I don’t care about what kind of glasses he’s wearing so the post already feels a bit weird to me but outside of that. If these are jokes and I hope these are jokes, they are jokes that you would make for a fictional character. Similar to “I wanna stick this fictional character in a blender”; but that’s obviously a joke because the character can’t actually get hurt. There’s no tangible effect if you shoved a fictional character in a blender. Except these are jokes about a real person; not a fictional character. There would be a tangible effect if someone were to actually bully Brighting for wearing women’s frames. Which is also stupid because even if he did wear mainly women’s frames, why does it matter? Is that really something that would make you feel less about a person? Hopefully not, because it really doesn't matter and I can't think of a way where it would. Also men can wear women’s frames as long as the frame size is right for his proportions, and most frames are labeled as ‘unisex’ so it matters even less. Imagining genuinely having the want to bully an older guy because he doesn't present himself in a completely masculine way, despite him wearing different frames in every other photo and not even wearing glasses in some. This is why I’m dead set on these being jokes because if they aren’t, as a gender-non-conforming person, I would be so mad. This is a weird thing to fixate on because glasses frames are unisex; so not even people who are weird about gender roles give a shit. Even my 70+ year old grandpa didn’t care when I was trying out new glasses. It really doesn’t fucking matter. Though on the plus side, the other half of the comments were like ‘Leave him alone. He just wants to and he can do what he wants”. Though there were a few people that called him ‘girly pop’ in Brighting’s defense, I'll get to that. It does show how easy it is for people on TikTok to think of reality similar to fiction; which makes some users say stupid shit about the real person. 

Some people are also way too comfortable when talking about Brighting, thinking of him as a best friend or an ‘icon’. That one person who called Brighting “girly pop” is a good example of how comfortable people can get when talking about him, despite not even sharing the same air as him. Like I see people respond to his tweets with “bestie” and “I love you”. It could be just me but it takes years for me to be best friends with someone and to tell them “I love you” is even longer. So the words and phrases I associate with long-lasting friendships are being used for someone who hasn’t even talked to you. If that isn’t parasocial, I don’t know what is. In one case I particularly remember, I saw a TikTok of someone thinking about Brighting as like an adoptive father figure. Basically wanting Brighting to adopt them. Which…I'm not gonna completely disregard the entire idea. I understand the want for a respectable father figure in your life; hell, the relationship with my dad is nonexistent. I can see how people think of him in a parental way and in most other positive ways. However, I worry for the people who attach themselves to Brighting in this one-sided way if they aren't careful. Especially if they perceive him as such a personal figure, like a best friend or family member. Because when you perceive someone in this regard, you're going to feel the same emotions that you would with actual friends or family members. This goes for positive  emotions, and unfortunately, this also goes for negative emotions. And don’t even get me started on perceiving him romantically or… suggestively because those emotions are even stronger. I’ve even seen people on TikTok openly say they’re “simps” for Kevan Brighting; the real guy that can easily see your comment and understand what you’re talking about by looking it up. 

The emotions towards the person are increased to such a strong and personal level, simply put. So when he disappoints you or does what you perceive as disappointing, it’s going to be more difficult to brush it off than if you thought of him as just a guy. As someone who had parasocial relationships like that at a young age, it’s very difficult to get over it and it’s even more difficult to recognize that you’re doing it. So the one thing I just need to say that you have to know is the famous person of your attachment  doesn't know you exist. Kevan Brighting is not your friend. That’s not his fault and that’s not your fault. He’s just not reasonably going to know every person that knows him, and he’s probably not gonna be on-board with every person’s perception of him. If I was in his position, I would not feel comfortable being regarded as a parental figure or a best friend by people that I don’t even know. But I’m saying this to myself as much as I’m saying this to other people because I make this mistake as well. I get too comfortable talking about people I know nothing about. I’m not innocent in this regard. I get too fixated over real life people sometimes. Despite me knowing about Brighting’s career in voicework and broadcasting, I really don’t know him personally. I don’t know his age, his height, his favorite fucking color and I certainly don’t know how he is as a person. I can make guesses but they are just guesses. Just because I have the opportunity to acknowledge him doesn’t mean he has the opportunity to acknowledge me, and that’s fine.

It’s fine to talk about how talented his voice acting is, or even how attractive/hot his voice is. Because he’s indeed talented and his voice does sound very attractive. But there’s gotta be a limit. I also understand thinking positively about him because he seems like a genuinely friendly and down-to-earth person. But notice how I said ‘seem’? That’s because, to reiterate, I don’t know him personally. He could be eating souls and I would be none the wiser (THAT IS A JOKE). Kevan Brighting is a real-life person with thoughts and feelings and his bounds should be respected. Don’t be like the people I just talked about. He’s talented as hell in voice acting, and that is enough for the community to talk about. It definitely was for me; like look at my “Narrator’s Voice” section again. Straight up 10 pages. Hell, I’m planning on writing something else about his other performances because his voice is just…too good. Yeah, that’s right! I’M NOT DONE!

(We Are Experiencing Some Technical Difficulties)

TL;DR: Don't be weird over real people

[librariansrose]

yo kuro theres some people harassing j's va over being nuetral on a ship, what u gotta say abt that? points to envyofsteel's post..

Genuinely not sure if you’re sending this to try and bait me to get involved in drama or something but I’ll answer anyway.

I think harassing the VA’s or any of staff is disrespectful and Dev would be ashamed of those doing so. Regardless of whatever the reason. Dev actually Has made direct confrontations towards the fans and same with their assistant Gen doing so as well.

Did everyone forget about the disrespect towards swords and some of the other VA’s by ripping and using their voices for ai? Or Dev’s assets? Or you know, maybe the fact people keep breaking the age limit rule Dev has on the series? Especially regarding discords?

Think that spells out my point.

Personally I care for maybe five or six drdt ships and everything else I have muted or if it makes me uncomfortable? It’s blocked. Yall should maybe try that too because in all honesty, bitching and moaning about things that bug you only serves to make Dev look bad and pushes the VA’s, Staff, etc away from the community as a whole. As opposed to blocking and muting which hurts absolutely no one at all. You also don’t waste energy on something that ultimately doesn’t matter more than more pressing issues on this planet. Like politics, school, work, national crisis, the list goes on.

Also I lowkey have the person you’re mentioning blocked since they helped with dogpiling me when I was enduring harassment and stalking before, so yea. Though I wont be going into this topic at all since I’ve finally found peace and people are leaving me alone as I’ve been requesting for years, so uh. Yea. I want people to respect that and no longer put things about it in my messages, inbox, server, whatever. Hatchet is buried and I have my peace, let it rot and fade my dude.

That being said. Maybe yall should stay out of my inbox about shit like this tbh?

I’m just a ghost writer who hosts events for the stuff I like. I deadass ignore or block everything else. Yall should really do that too.

Plus like. Her VA has endured harassment before this year and that is utterly unacceptable.

If I were Dev, I’d be embarrassed of all the fans going after my friends and staff.

But seeing as I just stick to my friends and ignore everything else, then of course I don’t get involved in any of that.

TDLR: You will not agree with everyone on this planet, but there are more important things to focus on. Use your block button, remove people from your space, don’t embarrass Dev and staff or hassle them. Don’t drag people outside the situation into drama when they’re dealing with their own lives. I literally hate Ace Markey and don’t give a fuck about him in general. I just support my friends who draw or make things about him because they work hard on their creations. That does not mean I myself, care about any of this.

Anyway XanTeruVid, Teruvid, XanRuko, XanVid, TeruNika, MaiDiana and LevRonika are awesome. All else I lowkey don’t care, have muted or in rare case’s I will write a commission or trade for a friend if I’m comfortable doing so.

That aside, everyone step outside for a few minutes. Take a few breaths, shut off your computers and maybe read a book or go work on a hobby you like.

Peace

[the8thsphynx]

@celestrian-gods

I’m sorry for tagging you like this, but I’m just throwing this in here in case any of my moots or followers want elaboration, too.

So the quickest and bluntest answer I can give is that etiquette has gone COMPLETELY out the window and now everything is a competition on who can be the biggest hater, and yes I DO blame Twitter culture.

Fanart is regularly stolen, uncredited, and slapped into someone’s Tiktok which gets 100x the engagement that the original artist got and when the artist is rightfully mad about it they get dogpiled and told they’re ‘taking it too seriously’ (as if tiktok doesnt pay people after a certain point). Let alone all of the AI bullshit and how AI ‘fanart’ is getting more engagement and praise for 1/100000th of the work and love.

Someone posts their fanwork that might not being the most liked quality or style (art, cosplay, doll, prop, keychains, etc…) and they suddenly become the hot new cringe meme and are driven to complete mental breakdowns for laughs.

Everyone is clambering over each other on harmless things being Harmful Bad Problematic and on actual harmful things (predatory community members, VAs, studios, homophobia and racism, etc…) being ‘Not That Deep’. Not to mention god forbid you like ANYTHING someone has to get in your space and let you know you’re wrong.

Now did the older generations of anime fans have issues?? Oh fuck yes, no one has ever been faultless. But there was at LEAST some kind of code of conduct and we all agreed to ‘Not Feed the Trolls’ by rewarding bad behavior with attention.

It feels like we’re at the point of No Return, too, because ANYONE’S attempt to defend someone getting harassed, call out, and rebuke any of they shitty behavior ends up just another cringe meme.

So yeah, the state of being an anime fan is absolutely Suck™️ right now.

At the risk of sounding like a grandpa, the new generation of anime fans sucks.

[akiharashizuka]

Yu-Gi-Oh Vrains episode 22 thoughts

Blackened Sun

The writers sure aren’t wasting any time, showing the eclipse right from the beginning. I thought it would last for longer, but it’s over before the opening plays. 

Since Kitamura was shown surprised by this incident, it wasn’t caused by SOL. Most likely. Of course, the main suspects are the Knights of Hanoi.

I liked that Ai mentioned how in the past eclipses were considered signs of disasters. I have to agree with Yūsaku, no matter how much time passes, there would still be people who believe in all kinds of superstitions.

If only Shima would start talking about the incident and stop bragging about how much knowledge he has...

I liked how Yūsaku just walked off, almost leaving him talking by himself xD

So, this “athnoer” (how do you even pronounce that? xD) incident became quite a spooky urban legend. And it’s very similar to what happened to Aoi.

That part when Yūsaku corrects shima though xD It took me a while to realize that he actually knew the info and acted confused in front of Shima because he was using the wrong term xD

Well, this footage already proves the differences between the urban legend and the actual incident. Clearly, the 12 AM part is to make things seem spookier. Also, the white hands-part reminded me of the infamous line from the Pokemon games: “That white hand on your shoulder...I’m only imagining it”. That was a very creepy part.

It was mentioned at some point that Anothers were considered to be people who gave up their real life and decided to live in Link Vrains. That theory is proven wrong by the footage Kusanagi provided. but I’m wondering what if there were people like that...Maybe the anime will touch this subject.

I’d like to add one more common trait to the ones Yūsaku mentioned: all of them seem to be old enough to be in highschool. Maybe even younger, but no more than middle school. Not sure if this was intentional...

I wonder if Makoto was expecting to be targeted, so he recorded what was happening to him in Link Vrains with hopes that whoever discovers him will notice that too. 

Hmm...can’t he just check if he has an AI or not? That would make things less troublesome...Especially that they narrowed people down to those who have the old Duel Disk, with no AI...

It was revealed in a summary that Gō will be involved in this incident and now we know why: because one of the victims happens to be his friend. 

Hmm, Genome mentioned that he dropped below the first 50 most popular duelists...I don’t think his previous rank was mentioned, but that’s quite a drop. I wonder what happened there. I doubt it’s only because he lost to Playmaker.

I also find it strange that Genome tried to recruit him.

Guess this part proves that whatever Kogami threw into that pit and Another are different things. It looks like revolver caused the incident in order to find out who Playmaker really is, before Kogami’s program activates. Also, I believe the eclipse is related to the program, rather than Another.

It’s good that we have a flashback about the two. It’s always nice to know more about characters and what drive them in general. Plus, Gō hasn’t made an appearance in a while, so it’s very welcomed. From this, I think none of them were involved in the Hanoi Project. 

Gō as a kid looks very weird to me, for some reason...

Whoa...That was quite tense. Until he realized that Yūsaku was wearing the same uniform. 

Hmm, so there are people who join Hanoi just to cause trouble. I thought they would have stricter conditions for new members.

We also get to take a look at other members, along with Genome who showed up earlier. It seems they have completely other goals, compared to the ones shown earlier harassing the duelists. 

I find their names interesting. Up util that Genome guy showed up , I thought that all Knights of Hanoi had names based on firearms, since it was pointed out that there is, apparently, one named Specter (also, “specter and “spectre”  are more or less the same thing...). But now I can go back to my initial thought that his name may refer to the supernatural apparition (ghost, phantom). As for these two, the blue-haired guy is named Faust. I think it might be a reference to the German fictional character who made a pact with the devil. As for the girl’s name, Vyra, I don’t know what it can come from...Maybe the wiki will discover something in a few days. Oh, and she’s voiced by Ami Koshimizu, a VA I like.

Considering that Gō disguised as a Knight of Hanoi in the past, I was almost sure it was him xD Also, how he talks and those fighting movements were pretty much hinting to that.

The reason why I didn’t talk about Genome’s name earlier was because he mentions it himself. so, yeah, it is a term used in genetics. 

Comparing DNA to programming...I always found the DNA structure much easier to understand than programming, but guess that’s because I only studied the basics of it...How the arrangement of genes determines traits is quite complex.

Ok, either he is bluffing or he did realize that Yūsaku is Playmaker from only that short interaction. If that’s the case, then I really want to know how exactly he came to this conclusion. Because no one else did. Not even Akira, who seems the smartest one in the bunch. 

So, that’s the Data Gale mentioned by Vyra. A gale is a very strong wind, but I don’t think it’s stronger than a storm. Oh well, there might be more to it than just make movement harder.

The thing I find the most interesting in the preview is Gō changing his style. Maybe that’s why he dropped in popularity. The transition from a style to another doesn’t always go smoothly.

[reseau-actu]

In a recent study

, Chinese and American researchers developed an artificial intelligence (AI) system using 1·3 million electronic health records from across China. Their system was capable of diagnosing diseases as accurately as an experienced paediatrician.

This was an amazing feat, considering the challenges involved. The human body is a complex piece of biological machinery and trying to diagnose our ailments is a monumental task, even for the most experienced doctors. To improve their diagnoses, doctors gather as much information as they can from their patients. This information comes from a wide range of sources, including genome-wide studies, demographics, doctor's notes, clinical images, laboratory results, genetic tests, or even from wearable sensors.

Today, databases of electronic health records from around the world hold such information for billions of patients. But, despite all the data and the best intentions, misdiagnoses happen more often than you might think. “Clinicians think and act in chaotic, time-pressured, clinical environments while dealing with uncertainty, which is integral to the diagnostic process”, Dr Hardeep Singh, a patient safety researcher at the Michael E DeBakey VA Medical Center and Baylor College of Medicine, in Houston, TX, USA, tells The Lancet Digital Health.

All this pressure in the fast-paced primary care environment can translate into diagnostic errors, especially when a doctor runs into problems gathering or interpreting information related to patient history, physical examinations, or tests, Singh explains. He adds, “Things are not always black-and-white when patients' symptoms and disease conditions are unfolding and dynamically evolving over time. Now add to this the currently immature and disconnected data information systems and you get the perfect storm of risk”.

A 2014 study

led by Singh found that at least one in 20 adults in the USA left the doctor's office with a misdiagnosis, which equates to 12 million people per year. Half of these misdiagnoses, the authors estimated, are potentially harmful.

One way to help physicians improve their diagnostic accuracy is by helping them analyse data more efficiently. With this goal in mind, and as computers become more powerful, scientists are looking at AI to design systems that can think like humans, only faster and more efficiently. AI approaches, such as machine learning, natural language processing, representation learning, and deep learning, aim to identify patterns in words, images, voice and video recordings, and other data in record time.

To develop an AI system to diagnose paediatric diseases, the Chinese and American team first created a dictionary with key words linking symptoms to a disease. This dictionary, which holds more than 1 million such key terms, was then used to train the AI system. “You can imagine that this requires both a huge dataset and a lot of human physician manpower to curate data—but once it is made, it can save a huge amount of work and has enormous impact to health-care delivery”, Prof Kang Zhang, lead author of the study, at the University of California, San Diego, CA, USA, tells The Lancet Digital Health.

This AI system was then used to analyse plain language inputs that would refer to symptoms with the goal of finding patterns that linked to a disease. To test their system, Zhang and his team used data from retrospective electronic medical records. “We tested diagnoses in 50 000 patients' EMR data as a validated cohort and additional 10 000 patients were tested when we compared accuracy of AI vs physicians” Zhang explains.

The results were nothing short of amazing, with a diagnostic accuracy of 95% for common ailments such as acute upper respiratory infection and sinusitis. The key to their success was in the amount and quality of the data available to train the AI algorithm and in the target population in whom it was tested; and here lays a potential limitation of this AI system. “AI can only work with diseases it has seen and trained before” Zhang says. So, the more data available for training, the more accurate the prediction the AI system can make, Zhang explains.

So far, the AI system designed by Zhang and his teams remains a proof of concept. A promising research experiment, not yet in use with actual patients. But, there are other AI systems that are already interacting with patients, such

as IBM's Watson

or

Babylon's AI chatbot

.

Babylon's AI system, for example, is currently being used by

over 3 000 000 users

from the UK and Rwanda. The system provides quick and easy-to-access advice for common ailments. But, how accurate is this advice?

In a non-peer-reviewed 2018 study

, Babylon researchers compared the performance of human physicians with Babylon's AI system at evaluating identical medical cases. According to their results, the AI system outperformed the average human doctor.

However,

in a 2018 correspondence

in

The Lancet

, Hamish Fraser, Associate Professor of Medical Science at Brown University, Providence, RI, USA, argued that Babylon's AI system still needed more testing before being able to prove it can perform well in real-world settings. For instance, Babylon's study did not use real patient data, but so-called patient vignettes, clinical examples of patient-related cases that are normally used for educational purposes. Other issues with Babylon's study, Fraser says, involves data input into the AI system, which was done by doctors, not patients (lay users), and that few statistical tests were done to assess the significance of their findings.

“The biggest problem at present with diagnosis programs for patients is the lack of rigorous, clinical evaluation. The studies to date have almost all been small and have not used real patient data so we don't know if a system like Babylon performs well with typical patients presenting with a wide range of clinical problems”, Fraser says.

Saurabh Johri, Chief Scientific Officer at Babylon, commented on this lack of rigorous clinical evaluations and says that studies are on their way. “We are keen to run robust, real-world studies on Babylon's AI that can be published in peer-reviewed journals”, he says. According to Johri, Babylon is currently working with academics to validate their AI technology. He added, “…we know new technology needs high quality evidence to earn people's trust”.

The question in people's mind is will doctors ever be replaced by computers? Experts agree that this is unlikely, at least under most circumstances. AI systems could be particularly useful in resource-limited areas or in emergency situations when a human doctor does not have the time or resources to reach a diagnosis. “I would envision AI can be used as a primary deployment to diagnose and triage diseases in rural or resource poor areas, for example, as an online system”, Zhang says. But even under these circumstances, Zhang argues that human doctors should still be involved and that “AI should be used primarily as a physician assistant to reduce overall burdens of physicians and it should not replace physicians”.

Taken together, the use of AI in health care heralds an exciting and promising new era in diagnosis, in which meaningful medical guidance can be made available to more people in need.

Article Info

Publication History

Identification

DOI: https://doi.org/10.1016/S2589-7500(19)30011-1

Copyright

© 2019 The Author(s). Published by Elsevier Ltd.

User License

Creative Commons Attribution – NonCommercial – NoDerivs (CC BY-NC-ND 4.0) |

How you can reuse

Permitted

For non-commercial purposes:

Read, print & download

Redistribute or republish the final article

Text & data mine

Translate the article (private use only, not for distribution)

Reuse portions or extracts from the article in other works

Not Permitted

Sell or re-use for commercial purposes

Distribute translations or adaptations of the article

Elsevier's open access license policy

ScienceDirect

Access this article on ScienceDirect

Karl GruberKarl GruberSearch for articles by this author Source: The Lancet Digital Health