Slashdot is powered by your submissions, so send in your scoop

 



Forgot your password?
typodupeerror
×
AI Iphone

iPhones Will Be Able To Speak in Your Voice With 15 Minutes of Training (theverge.com) 63

Apple today previewed a bundle of new features designed for cognitive, vision, hearing, and mobility accessibility. That includes a new Personal Voice feature for people who may lose their ability to speak, allowing them to create "a synthesized voice that sounds like them" to talk with friends or family members. From a report: According to Apple, users can create a Personal Voice by reading a set of text prompts aloud for a total of 15 minutes of audio on the iPhone or iPad. Since the feature integrates with Live Speech, users can then type what they want to say and have their Personal Voice read it to whomever they want to talk to. Apple says the feature uses "on-device machine learning to keep users' information private and secure."

Additionally, Apple is introducing streamlined versions of its core apps as part of a feature called Assistive Access meant to support users with cognitive disabilities. The feature is designed to "distill apps and experiences to their essential features in order to lighten cognitive load." That includes a combined version of Phone and FaceTime as well as modified versions of the Messages, Camera, Photos, and Music apps that feature high contrast buttons, large text labels, and additional accessibility tools.

This discussion has been archived. No new comments can be posted.

iPhones Will Be Able To Speak in Your Voice With 15 Minutes of Training

Comments Filter:
  • Unfortunately (Score:5, Insightful)

    by Viol8 ( 599362 ) on Tuesday May 16, 2023 @11:02AM (#63526015) Homepage

    Scammers will be rubbing their hands with glee here. If people are fooled by text messages supposedly coming from family members in trouble and needing money , imagine getting a call from what sounds like them asking for it.

    • You have to get someone to read through a series of text in order to re-create their voice...

      Plenty more deep-fake audio engines being built that will not have such rigid training, Apple's is probably the last one you need to worry about.

      But yeah, deep fake phishing stuff is coming, probably already here...

      • Hmm...I wonder if Apple got this idea from one of the Mission Impossible movies...

        I think Tom Cruise made Phillip Seymour Hoffman read something....and then could copy his voice?

      • You have to get someone to read through a series of text in order to re-create their voice...

        Which doesn't make any sense if you've already lost your voice.

        What are we supposed to do? Read the texts and store the voice data in the cloud just in case we lose our voices via some future random event?

        That stinks of rat to me - almost as if some big company wants some free voice training data.

        • Re: (Score:1, Informative)

          by Anonymous Coward

          You have to get someone to read through a series of text in order to re-create their voice...

          Which doesn't make any sense if you've already lost your voice.

          What are we supposed to do? Read the texts and store the voice data in the cloud just in case we lose our voices via some future random event?

          "Personal Voice, on the other hand, is a way for people who are at risk of losing their ability to speak to create and save a voice that sounds like them. Apple says it’s designed for people at risk of losing their ability to speak, such as those with a recent diagnosis of ALS."

          That stinks of rat to me - almost as if some big company wants some free voice training data.

          Right, because audio of people speaking is so hard to come by otherwise.

          • >Right, because audio of people speaking is so hard to come by otherwise.

            That and Apple specifically aims for local processing as much as possible.

            • That and Apple specifically aims for local processing as much as possible.

              "As much as possible"?

              Translation: So they don't have to pay to store any data they can't profit from.

            • That and Apple specifically aims for local processing as much as possible.

              FWIW: This HAS to be uploaded to Apple. Would you want to lose your voice just because you dropped your iPhone and broke it?

          • Right, because audio of people speaking is so hard to come by otherwise.

            Not everyone posts themselves yakking on social media twelve times a day.

          • Right, because audio of people speaking is so hard to come by otherwise.

            Good luck finding some of mine.

    • by Tablizer ( 95088 )

      While in early stages of Alzheimer's, my dad was scammed out of about $10k by a human impersonating a "desperate" relative. AI will just make such even easier.

      I actually don't think the scammers targeted my relative's voice. Instead they probably cold-call a list of elderly, purchased on the underground market, and if the voice sounded familiar, the callie talked longer, trying to get more info. For example, the callie might ask, "Fred? Is that you?" The scammers then call few months later using those clue

  • by Joe_Dragon ( 2206452 ) on Tuesday May 16, 2023 @11:05AM (#63526027)

    Hi, my name is Werner Brandes. My voice is my passport. Verify Me

    • I came here to post this for my fellow Olds.
      • Are you suggesting anybody under the age of 44 uses Slashdot? This place is basically an AARP convention.

        • Oi. Not all of us are old enough for an AARP convention. Based on the SAO summary from last year [slashdot.org], I'd say some of the writers for /. aren't either. Although looking at the comments for that "story" clearly indicates most people on /. are well above 44....(and that /. needs at least two days to get a story based on a fictional event known decades in advance out. :P)
    • by hawk ( 1151 )

      Oh yeah?

      Well *my* name is Inigo Montoyo.

      You killed my father.

      Prepare to die!

  • by Anonymous Coward

    That includes a new Personal Voice feature for people who may lose their ability to speak, allowing them to create "a synthesized voice that sounds like them" to talk with friends or family members. From a report:
    According to Apple, users can create a Personal Voice by reading a set of text prompts aloud for a total of 15 minutes of audio on the iPhone or iPad.

    So anyone who can't speak just has to read a set of text prompts aloud... wait a minute, I think I see a problem here.

  • Of course, this will not be abused. Taking a widespread consumer device and having it sound like a politician or celebrity is not tempting at all. "Hi, Mr. Smith. I'm Joe Biden. Let's go Brandon. *CLICK*" "Hi, Mr. Smith. This is Donald J. Trump. The Donald. *click*" "Hi, Mr. Smith. This is Bill Clinton. Some people call me Slick Willy. I'm calling to let you know that in these tough economic times, the president's tool is so incomprehensively large. *click*" I mean, what could go wrong?
    • I don't think that first example would work, Biden seems to be owning it, which is unusually clever for a Democrat.

      I'm already getting a *ton* of Youtube scam adverts with Joe Rogan and Elon Musk and other right wing stars shilling for crypto coins and tactical fanny packs and other nonsense though.

      During the last election one of the major problems the Republican party had was there was so much grifting going on they couldn't raise money for smaller races that the big money guys weren't backing. The
      • I don't think that first example would work, Biden seems to be owning it, which is unusually clever for a Democrat.

        Well, to be fair...Biden can't string together more than 2-3 words in a row in a coherent manner...nowhere near enough for this Apple device to learn anything.

        Kamala is worse...but she goes the opposite direction...many, many, many, many words strung together in a "word salad" that still isn't coherent or meaningful.

        • “Look, having nuclear — my uncle was a great professor and scientist and engineer, Dr. John Trump at MIT; good genes, very good genes, OK, very smart, the Wharton School of Finance, very good, very smart — you know, if you’re a conservative Republican, if I were a liberal, if, like, OK, if I ran as a liberal Democrat, they would say I’m one of the smartest people anywhere in the world — it’s true! — but when you’re a conservative Republican they try

          • Never said Trump was any better a speaker....

            Although even though almost the same age, I'd say Don was more coherent than Joe....

            I've not seen Don get lost on a stage wondering around trying to shake hands with "the air"....

            He also doesn't still get lost even when his staff hands him notes that include such things as "sit down"...this means you...etc.

            But I digress....I don't really want either one of them, but if it came down to it....Joe would be worse.

            A big reason for that is, I don't trust Joe would

        • Well, to be fair...Biden can't string together more than 2-3 words in a row in a coherent manner...nowhere near enough for this Apple device to learn anything

          The more you right wingers make this claim the more you actually underestimate how cognizant our current POTUS is. The guy that outmaneuvered the GOP during the State of the Union address ain't incompetent.

          In case you missed it, Biden has an actual speech impediment (stuttering) that he's had to deal with his entire life, which may explain why you think he's incompetent due to how he speaks.

          And I'd much rather have an old POTUS who is in Biden's physical and mental condition instead of tRump's.

          • In case you missed it, Biden has an actual speech impediment (stuttering) that he's had to deal with his entire life,

            Funny...I saw absolutely NO sign of any impediment from Biden during his younger days as senator and even while VP for Obama.

            The only thing he had was an incredible capability for gaffes....

            It's easy to go back and see interviews with him....and his time in the senate and nope...no sign of any discernible impediment.

            However, now..? It isn't an impediment...it is him starting to talk and

    • For one, this ship has already sailed. Second, I am sure that this will only be enabled if the voice training prompts are completed on the device in a timely manner. If so, you would have to pre-record all the answers to the prompts and then play them back as they were asked by the device.

  • I got three words for their training algorithm. OH, HELL, NO!
  • Ripe for abuse (Score:5, Interesting)

    by brunes69 ( 86786 ) <slashdot@nOSpam.keirstead.org> on Tuesday May 16, 2023 @11:19AM (#63526065)

    Step 1: Find video clips of someone speaking online

    Step 2: Train your iPhone to speak like them

    Step 3: Call their loved ones and convince them to give you private information to steal their identity or blackmail them or any number of other things.

    Yes, you could do this today with existing technology, but this will lower the barrier to an unacceptable level.

    • This requires matching prompts to speak specific words. So unless you're going to go through and find clips of every phrase or word required for training, you can't simply grab a video of a celebrity and be ready to go.

      • Never underestimate thieves.
        • It's a bit of a Hollywood plan, one that anybody with the competence and resources to pull off would not even attempt. There are far simpler methods. It'd be easier to kidnap the person.

  • If you want your iPhone to read a web page, the only way to do it is to select all the text of a page (which is hard to do if it is very long) and have it speak that. There is no option to simply select âoeRead Pageâ which is a shame.
  • ...nobody I know likes their voice outside of their head.

  • not useful unless you can type, and typing on an iphone truly sucks. this will only be useful for folks using keyboards. intersection of fast thumb typers and voice sufferers is empty.

    • ...and typing on an iphone truly sucks.

      How exactly is this? I don't find typing on my iPhone to be difficult at all.

      How is typing on an iPhone any different than typing on any other smart phone (without a physical keyboard, which pretty much none of them have anymore)?

  • That's... wonderful, isn't it!

  • by VeryFluffyBunny ( 5037285 ) on Tuesday May 16, 2023 @12:39PM (#63526291)
    (Verse 1)
    Yo, listen up, I got a story to tell,
    'Bout iPhone users, man, it ain't goin' well.
    We love our iPhones, they're sleek and they're fly,
    But there's one thing that's got us askin' why.

    We spend our hard-earned cash on these devices,
    Dreamin' 'bout customizing our vocal devices.
    But Apple says, "Hold up, you can't choose your voice,
    We'll decide for you, it ain't your choice."

    (Chorus)
    iPhone users, we're feelin' the blues,
    Locked in a box, can't pick the voice we use.
    Celebrity voices, oh, they'd be a blast,
    But Apple's sayin', "Sorry, this ain't happenin' fast."
    (Verse 2)
    We wanna wake up to Snoop Dogg's smooth flow,
    Or have Samuel L. Jackson yellin', "Let's go!"
    But Apple's got us trapped in their walled garden,
    No freedom of voice, it's beggin' for pardon.

    We see those ads, the celebs talkin' the talk,
    But we can't rock their voices, it's a cruel block.
    Hey Apple, why can't we live out our dream?
    Train our iPhones to sound like the A-list team.

    (Verse 3)
    We're loyal customers, we deserve some choice,
    Let us train our iPhones with a celebrity voice.
    Morgan Freeman narratin' our daily grind,
    Or Cardi B droppin' one-liners, one of a kind.

    Apple, listen up, we're beggin' you, please,
    Open up the gates, let our voices be free.
    We're ready to pay, we're ready to invest,
    In a world where our iPhones truly impress.

    (Bridge)
    To all my iPhone users, let's raise our voice,
    Demandin' the freedom to make our own choice.
    Apple, don't deny us this opportunity,
    Let us train our iPhones with celebrity quality.

    (Outro)
    So Apple, hear our plea, it's time to evolve,
    Let us choose the voices that make us revolve.
    Give us the power, let our iPhones shine,
    With celebrity voices, it's our time to define.
  • Will this thing imitate what I know my voice sounds like, or will it sound like the horribly distorted nasally version of my voice that consistently shows up in all audio and video recordings of me?

    I imagine a lot of people trying to use this feature will re-train it a few times because they think there's no way they sound like that.

  • the last thing i'd like to hear would be my own voice. why not some other, noble voices? like Arnold's?
  • Perhaps this will stop companies from using the voice verify technology that has started to pop up on some phone systems. Or it could have the opposite effect. I guess we'll see.
  • Its getting even more faker. Maybe they fix the whole fakeness but will it be 10-20 years of it? Wonder what a society will be like after that amount of experience. Either way its gonna be weird.
  • Will it be able to swear the wallpaper off of a wall?

"Being against torture ought to be sort of a multipartisan thing." -- Karl Lehenbauer, as amended by Jeff Daiell, a Libertarian

Working...