Latest / 6 Ps in AI Pods (AI6P) / AI Ventriloquism: Send in the clones [Unfair use? series, Part 3] 🗣️

Transcript

AI Ventriloquism: Send in the clones [Unfair use? series, Part 3] 🗣️ · 6 Ps in AI Pods (AI6P)

• 0:00
  SPEAKER_00: Hi everyone, this is Karen Smiley.
• 0:02
  SPEAKER_00: Welcome to 6 Ps in AI Pods on Substack.
• 0:07
  SPEAKER_00: AI ventriloquism, send in the clones.
• 0:10
  SPEAKER_00: Discoveries on how common voice cloning tools are and the lack of attention to ethics by most of them.
• 0:16
  SPEAKER_00: This is a bonus article in our series on infer use and ethics of generative AI for music.
• 0:22
  SPEAKER_00: While doing my diligence on the broad landscape of GenAI-based tools for our Part 3 article,
• 0:27
  SPEAKER_00: I came across a whole bunch of irresponsible voice cloning tools.
• 0:31
  SPEAKER_00: Here's a summary of what I found, with recommendations for ethical ways to meet voice cloning needs.
• 0:38
  SPEAKER_00: As a reminder,
• 0:38
  SPEAKER_00: this article is not a substitute for legal advice,
• 0:40
  SPEAKER_00: and it is meant for general information only.
• 0:44
  SPEAKER_00: What are voice clones and vocal covers?
• 0:48
  SPEAKER_00: AI-based voice clone tools train individual voice models on a specific speaking voice.
• 0:55
  SPEAKER_00: Then the tool applies a voice model to make that voice speak words or phrases which
• 1:00
  SPEAKER_00: the speaker most likely never said.
• 1:03
  SPEAKER_00: If the tool user is using a voice model that's not their own voice,
• 1:07
  SPEAKER_00: the tool user can create a recording of the voice owner saying words without their consent.
• 1:13
  SPEAKER_00: This is AI-based ventriloquism with real people instead of puppets or dummies.
• 1:19
  SPEAKER_00: Voice cloning can make it seem that people spoke words,
• 1:22
  SPEAKER_00: phrases,
• 1:22
  SPEAKER_00: or sentences,
• 1:23
  SPEAKER_00: et cetera,
• 1:23
  SPEAKER_00: that they never did.
• 1:26
  SPEAKER_00: There are innocuous applications,
• 1:28
  SPEAKER_00: such as text-to-speech for accessibility,
• 1:31
  SPEAKER_00: or a grandparent singing a lullaby or reading a bedtime story.
• 1:36
  SPEAKER_00: But the same tools can just as easily be used to make recordings of people saying
• 1:40
  SPEAKER_00: lies or sentiments they might vehemently disagree with.
• 1:44
  SPEAKER_00: It's putting words into virtual clones of people's mouths, often without their consent.
• 1:50
  SPEAKER_00: If you've heard about scams using deep fakes of voice recordings, these tools create them.
• 1:57
  SPEAKER_00: Creating and using fake recordings without proper licensing,
• 2:00
  SPEAKER_00: consent,
• 2:01
  SPEAKER_00: credit,
• 2:01
  SPEAKER_00: and compensation from the voice owner is highly unethical,
• 2:06
  SPEAKER_00: period,
• 2:07
  SPEAKER_00: and potentially extremely damaging.
• 2:09
  SPEAKER_00: And that's what many of these tools enable.
• 2:13
  SPEAKER_00: Similarly, vocal cover tools are style transfer tools based on generative AI.
• 2:18
  SPEAKER_00: They apply a model trained on a specific singing or speaking voice to a song by a different artist.
• 2:25
  SPEAKER_00: They don't generate new melodies, but they do generate new recordings with a substitute vocal track.
• 2:32
  SPEAKER_00: Essentially,
• 2:33
  SPEAKER_00: they then use AI to make a recording that sounds like the speaker or vocalist is
• 2:38
  SPEAKER_00: singing a different song.
• 2:41
  SPEAKER_00: There are two ethical issues to consider with vocal covers.
• 2:45
  SPEAKER_00: One,
• 2:45
  SPEAKER_00: the rights of the speaker or singer whose voice is being applied to the song,
• 2:49
  SPEAKER_00: and two,
• 2:49
  SPEAKER_00: the rights of the owner of the song.
• 2:53
  SPEAKER_00: The Rise of AI Voice Cloning Tools To be clear, it's no surprise that these voice cloning tools exist.
• 3:00
  SPEAKER_00: They're not new, and many people have raised alarms about the potential for scams and disinformation.
• 3:06
  SPEAKER_00: What distressed me was realizing that the unethical use of voice cloning is
• 3:10
  SPEAKER_00: seemingly so unquestioned by so many sites and reviewers and users.
• 3:15
  SPEAKER_00: In my research,
• 3:16
  SPEAKER_00: I found many articles about these tools that don't even mention ethical concerns,
• 3:20
  SPEAKER_00: and only a few that do mention them.
• 3:24
  SPEAKER_00: Given that surveys have shown that most Americans feel musicians and other artists
• 3:28
  SPEAKER_00: should be paid for their work,
• 3:30
  SPEAKER_00: this nonchalance about voice cloning feels like a disconnect.
• 3:34
  SPEAKER_00: And one doesn't even need to venture into dark corners of the web to find these voice cloning tools.
• 3:38
  SPEAKER_00: They're right out there in public.
• 3:41
  SPEAKER_00: It's one thing for artists to develop and use models of their own voice.
• 3:45
  SPEAKER_00: A few have.
• 3:46
  SPEAKER_00: Artist Holly Herndon was one of the first to break ground with AI.
• 3:50
  SPEAKER_00: Her 2021 release of her Holly Plus AI voice model,
• 3:54
  SPEAKER_00: a digital vocal twin,
• 3:56
  SPEAKER_00: will,
• 3:56
  SPEAKER_00: quote,
• 3:57
  SPEAKER_00: allow for others to make artwork with my voice and will distribute ownership of my
• 4:01
  SPEAKER_00: digital likeness through the creation of the Holly Plus DAO.
• 4:06
  SPEAKER_00: Grimes also developed and released an elf.tech tool to allow others to use her voice,
• 4:11
  SPEAKER_00: which she encourages people to use under 50-50 royalty split terms.
• 4:18
  SPEAKER_00: Of the tools not developed directly by artists like Herndon and Grimes,
• 4:21
  SPEAKER_00: only a few passed the ethical smell test.
• 4:25
  SPEAKER_00: For most AI-driven vocal cover tools I've looked at so far with built-in voice models,
• 4:30
  SPEAKER_00: I have yet to see any indication that the artist whose voice is modeled,
• 4:34
  SPEAKER_00: or the composer,
• 4:35
  SPEAKER_00: lyricist,
• 4:36
  SPEAKER_00: or instrument performers of the covered song,
• 4:38
  SPEAKER_00: receives any compensation or is given consent.
• 4:43
  SPEAKER_00: A handful of responsible options.
• 4:47
  SPEAKER_00: Relatively little attention has been given in reviews of these tools to ethical considerations.
• 4:52
  SPEAKER_00: One exception from December 2023 called out AI modeling websites are making money
• 4:58
  SPEAKER_00: selling the usage of these models of famous singers,
• 5:00
  SPEAKER_00: and very few of these websites were talking about these artists getting a cut of
• 5:04
  SPEAKER_00: the profits.
• 5:05
  SPEAKER_00: Of those we've tested,
• 5:06
  SPEAKER_00: only a few,
• 5:07
  SPEAKER_00: for instance,
• 5:08
  SPEAKER_00: Kits and VoiceSwap,
• 5:10
  SPEAKER_00: mention royalty splits with the original artist's affiliates,
• 5:13
  SPEAKER_00: who its AI algorithms are modeled on.
• 5:16
  SPEAKER_00: The website also lets these artists have the final say on commercial releases.
• 5:22
  SPEAKER_00: In searching for tools and reviews, I found several lists.
• 5:25
  SPEAKER_00: A few tool offerings appear to use AI responsibly.
• 5:29
  SPEAKER_00: 11labs, kits.ai.com,
• 5:32
  SPEAKER_00: Revocalize, Speechify, and VoiceSwap.
• 5:36
  SPEAKER_00: Let's summarize those first.
• 5:40
  SPEAKER_00: Eleven Labs offers the option to create a model of your own voice,
• 5:44
  SPEAKER_00: possibly taking some hours to render,
• 5:47
  SPEAKER_00: with some interesting technical precautions to try to ensure that it's really your
• 5:50
  SPEAKER_00: own voice that you're modeling.
• 5:53
  SPEAKER_00: They offer over 1200 voices across 29 languages and provide ethics resources on their homepage.
• 6:01
  SPEAKER_00: Kits.ai.
• 6:03
  SPEAKER_00: Their slogan,
• 6:05
  SPEAKER_00: clone voices,
• 6:06
  SPEAKER_00: sing like anyone,
• 6:07
  SPEAKER_00: play any instrument 100% royalty free,
• 6:10
  SPEAKER_00: is a bit concerning at first blush.
• 6:13
  SPEAKER_00: But their ethics page indicates they are trying to act responsibly.
• 6:19
  SPEAKER_00: They refer to responsible data set sourcing, fair artist compensation, and empowering creators.
• 6:27
  SPEAKER_00: Revocalize offers a small number of officially licensed voice models,
• 6:32
  SPEAKER_00: but the focus is on creating new models not using famous voices.
• 6:37
  SPEAKER_00: Their offering allows you to clone, protect, and create unique vocal tracks in any voice.
• 6:44
  SPEAKER_00: VoiceSwap offers session singers and a roster of exclusive compensated artists
• 6:51
  SPEAKER_00: whose voices are available for use through their tool.
• 6:54
  SPEAKER_00: The artist list on the site shows 16 names as of this writing.
• 6:59
  SPEAKER_00: Their FAQ asserts that as of July 2023,
• 7:02
  SPEAKER_00: VoiceSwap is the only platform that works legally with famous artists.
• 7:08
  SPEAKER_00: Speechify, which is not for songs, offers voices of celebrities they have partnered with, e.g.
• 7:14
  SPEAKER_00: Gwyneth Paltrow and Snoop Dogg, and they have precautions to ensure that only live voices are cloned.
• 7:21
  SPEAKER_00: Ethical voice cloning and text-to-speech tools have clear use cases that can add value.
• 7:28
  SPEAKER_00: In addition to the accessibility features Speechify promotes,
• 7:32
  SPEAKER_00: think how much time it might save a book author to AI generate an audiobook version
• 7:36
  SPEAKER_00: in their own voice,
• 7:38
  SPEAKER_00: or a newsletter author to generate a podcast recording to help them reach new listeners.
• 7:44
  SPEAKER_00: Did I miss any voice cloning or vocal cover tools,
• 7:47
  SPEAKER_00: whether Gen-AI based or not,
• 7:49
  SPEAKER_00: that you consider ethical?
• 7:51
  SPEAKER_00: Let me know.
• 7:55
  SPEAKER_00: So now, lots of irresponsible options.
• 7:57
  SPEAKER_00: The few sites listed above that let you create voice models ensure it's truly your own voice, not e.g.
• 8:03
  SPEAKER_00: the voice of someone for whom you have a recording, but not their consent.
• 8:06
  SPEAKER_00: This use seems okay, ethically and legally.
• 8:10
  SPEAKER_00: But creating a unique model of the user's voice is not the focus of most of the voice cloning tools.
• 8:16
  SPEAKER_00: Note, some of these tools offer features besides voice cloning.
• 8:20
  SPEAKER_00: Most other examples in this space don't appear to care about responsible and ethical uses of AI.
• 8:26
  SPEAKER_00: The main marketing for these tools promotes their pre-built voice models of people or characters.
• 8:33
  SPEAKER_00: If a site has no FAQ or policy statement on how they train their voice models,
• 8:38
  SPEAKER_00: or it doesn't say that the voice training data has been licensed,
• 8:42
  SPEAKER_00: and there's no publicity about them successfully negotiating with an artist to use
• 8:45
  SPEAKER_00: their voice,
• 8:47
  SPEAKER_00: I assume the voices aren't licensed and are being used in the cloning tool without consent.
• 8:53
  SPEAKER_00: To name just a few celebrities whose voices are offered via these tools,
• 8:56
  SPEAKER_00: and there are thousands,
• 8:58
  SPEAKER_00: Taylor Swift,
• 8:59
  SPEAKER_00: Michael Jackson,
• 9:00
  SPEAKER_00: Donald Trump and Joe Biden,
• 9:02
  SPEAKER_00: Narendra Modi,
• 9:03
  SPEAKER_00: Morgan Freeman,
• 9:04
  SPEAKER_00: Messi,
• 9:04
  SPEAKER_00: Ronaldo,
• 9:05
  SPEAKER_00: Freddie Mercury,
• 9:06
  SPEAKER_00: Donald Duck,
• 9:07
  SPEAKER_00: and Spongebob.
• 9:09
  SPEAKER_00: It is unreasonable to believe that all of these thousands of people or their
• 9:12
  SPEAKER_00: estates or character owners have consented to these uses of the voice they own.
• 9:18
  SPEAKER_00: Examples of such tools include Voiceify, which is now Jammable,
• 9:22
  SPEAKER_00: Find shares Singify and FindVoice,
• 9:25
  SPEAKER_00: Vocloner,
• 9:26
  SPEAKER_00: MusicFi,
• 9:28
  SPEAKER_00: UberDuck,
• 9:29
  SPEAKER_00: FakeYou,
• 9:30
  SPEAKER_00: Malows,
• 9:31
  SPEAKER_00: and Cover.ai,
• 9:33
  SPEAKER_00: which says it's for personal and parody use only.
• 9:37
  SPEAKER_00: Fake You doesn't have an ethics page or relevant information.
• 9:40
  SPEAKER_00: They say only that their mission is to empower anyone to create full-length,
• 9:44
  SPEAKER_00: feature-length content from home without institutional capital,
• 9:48
  SPEAKER_00: large teams,
• 9:48
  SPEAKER_00: huge amounts of time,
• 9:49
  SPEAKER_00: or deep reservoirs of highly specialized talent.
• 9:52
  SPEAKER_00: As of this writing,
• 9:53
  SPEAKER_00: they offer 3,966 voices for text-to-speech,
• 9:57
  SPEAKER_00: saying,
• 9:57
  SPEAKER_00: "...turn text into your favorite character's speaking voice,
• 10:00
  SPEAKER_00: including separate speaking and singing voices for Michael Jackson."
• 10:05
  SPEAKER_00: For voice-to-voice, they offer 8,617 voices.
• 10:10
  SPEAKER_00: They also now have a beta voice designer feature.
• 10:14
  SPEAKER_00: All of these seem ethically shady.
• 10:17
  SPEAKER_00: Vocloner allows users to
• 10:19
  SPEAKER_00: clone the voice of anyone in seconds with just need one audio file or the voice you
• 10:25
  SPEAKER_00: want to clone,
• 10:26
  SPEAKER_00: upload a sample audio file and enter the text you would like the voice to say.
• 10:30
  SPEAKER_00: There is zero information about ethical use.
• 10:34
  SPEAKER_00: They also now have an AI music generator in beta.
• 10:37
  SPEAKER_00: The homepage for Lalau's claims,
• 10:39
  SPEAKER_00: we ensure our service is compliant and legal and our AI voices are only trained
• 10:42
  SPEAKER_00: with legal material,
• 10:44
  SPEAKER_00: but they don't back that up with any details.
• 10:46
  SPEAKER_00: The voices they claim to have include Paul McCartney and John Lennon.
• 10:51
  SPEAKER_00: Online reviews for Lalau's don't address ethics or even this one,
• 10:54
  SPEAKER_00: which identifies other creative perils of tools like Lalau's.
• 10:59
  SPEAKER_00: A Voiceify user can select a voice model, pair it with a YouTube URL of any song,
• 11:05
  SPEAKER_00: and get a cover of that song with that voice model.
• 11:09
  SPEAKER_00: Voiceify has quickly expanded its pool of AI vocal models from 800 plus to 22,000 plus,
• 11:16
  SPEAKER_00: including,
• 11:16
  SPEAKER_00: quote,
• 11:17
  SPEAKER_00: voices of famous musicians,
• 11:19
  SPEAKER_00: politicians,
• 11:19
  SPEAKER_00: and cartoon or game characters.
• 11:22
  SPEAKER_00: There's no information on their site about ethics, data sourcing, consent, or compensation.
• 11:28
  SPEAKER_00: And this supposedly comprehensive,
• 11:30
  SPEAKER_00: damnable AI review,
• 11:31
  SPEAKER_00: everything you should know article from March 2024 doesn't address these ethical
• 11:36
  SPEAKER_00: concerns at all.
• 11:39
  SPEAKER_00: The FAQ for Singify AI Voice Cover response to a question about legality is,
• 11:44
  SPEAKER_00: The legality of AI-generated song covers depends on factors like intended use,
• 11:49
  SPEAKER_00: copyright laws,
• 11:50
  SPEAKER_00: and permissions.
• 11:51
  SPEAKER_00: Commercial use may require authorization from the original copyright holder,
• 11:56
  SPEAKER_00: while personal use may fall under fair use exemptions in some jurisdictions.
• 12:01
  SPEAKER_00: They do not address licensing or data sourcing.
• 12:04
  SPEAKER_00: FindVoice offers to quickly convert text or audio into the voice of your favorite character,
• 12:08
  SPEAKER_00: and add pauses,
• 12:10
  SPEAKER_00: emphasis,
• 12:10
  SPEAKER_00: and even unique personality,
• 12:12
  SPEAKER_00: and offers no information about consent from the favorite characters or other
• 12:16
  SPEAKER_00: voices they offer.
• 12:18
  SPEAKER_00: MusicFi appears to offer ability to use your own voice or one of their copyright-free pre-trained voices.
• 12:24
  SPEAKER_00: That seems okay.
• 12:26
  SPEAKER_00: But farther down the homepage,
• 12:27
  SPEAKER_00: they offer parody voices such as Barack Obama,
• 12:31
  SPEAKER_00: Bart Simpson,
• 12:32
  SPEAKER_00: SpongeBob SquarePants,
• 12:33
  SPEAKER_00: Mickey Mouse,
• 12:34
  SPEAKER_00: and Shrek.
• 12:35
  SPEAKER_00: Another page claims they have 1,000 plus celebrity voices.
• 12:39
  SPEAKER_00: It's pretty unlikely that all of those celebrity voice owners consented to this.
• 12:45
  SPEAKER_00: And MusicFi admits that its service raises ethical concerns when using famous voices,
• 12:49
  SPEAKER_00: but states,
• 12:51
  SPEAKER_00: you can freely use AI voice calling,
• 12:52
  SPEAKER_00: but it's important to exercise responsibility.
• 12:56
  SPEAKER_00: Musicify provides a blog page describing the permissions needed to cover a song,
• 13:00
  SPEAKER_00: but with no acknowledgement of the contribution of their tool to unethical or
• 13:03
  SPEAKER_00: illegal use.
• 13:06
  SPEAKER_00: Due to lawsuits from Universal Music and others,
• 13:10
  SPEAKER_00: Uberduck's former voice library dropped from 5,000-plus to 227 text-to-speech voices,
• 13:17
  SPEAKER_00: 15 AI vocal voices,
• 13:20
  SPEAKER_00: and one rap voice with several backing tracks as of March 25th.
• 13:25
  SPEAKER_00: At this writing, UberDuck currently offers 213 voices for text-to-speech and only 11 for voice-to-voice.
• 13:33
  SPEAKER_00: Two of those belong to Grimes.
• 13:36
  SPEAKER_00: These voice cloning tools and the unlicensed celebrity voices they provide make it
• 13:40
  SPEAKER_00: impossible for us to believe that any politician or public figure actually said the
• 13:44
  SPEAKER_00: words on any voice recording we may hear.
• 13:47
  SPEAKER_00: This is a clear path to election interference, folks, in the U.S.
• 13:51
  SPEAKER_00: and worldwide.
• 13:52
  SPEAKER_00: It's already been happening for a few years.
• 13:54
  SPEAKER_00: And as the tools get better at sounding more realistic and human, the problem is only going to get worse.
• 14:02
  SPEAKER_00: The real existential risk is that we all become mad because we cannot believe what
• 14:07
  SPEAKER_00: we are seeing,
• 14:07
  SPEAKER_00: hearing,
• 14:08
  SPEAKER_00: or reading.
• 14:09
  SPEAKER_00: That's from Carme Artega, Spain's Secretary of State for AI and Digitalization.
• 14:14
  SPEAKER_00: What we can do about it.
• 14:16
  SPEAKER_00: Laws.
• 14:18
  SPEAKER_00: Unauthorized voice cloning isn't illegal everywhere yet.
• 14:21
  SPEAKER_00: Hopefully it's only a matter of time.
• 14:23
  SPEAKER_00: Although our voices aren't covered by copyright, most countries recognize a person's right of publicity.
• 14:30
  SPEAKER_00: Right of publicity allows a person to control commercial use of some aspects of
• 14:33
  SPEAKER_00: their identity,
• 14:35
  SPEAKER_00: such as their name,
• 14:36
  SPEAKER_00: image,
• 14:36
  SPEAKER_00: and likeness,
• 14:37
  SPEAKER_00: NIL,
• 14:38
  SPEAKER_00: where voice is considered as part of one's likeness.
• 14:42
  SPEAKER_00: In the US,
• 14:42
  SPEAKER_00: if you've heard about NIL funds motivating NCAA athletes to transfer schools to get
• 14:48
  SPEAKER_00: more favorable deals,
• 14:49
  SPEAKER_00: yes,
• 14:50
  SPEAKER_00: this is the same NIL concept.
• 14:54
  SPEAKER_00: In the US, like in so many other areas of law, we have a patchwork of state laws.
• 14:59
  SPEAKER_00: The right of publicity isn't yet guaranteed at the federal level.
• 15:03
  SPEAKER_00: The U.S.
• 15:04
  SPEAKER_00: No AI Fraud Act, introduced in January, isn't universally agreed on.
• 15:09
  SPEAKER_00: For instance,
• 15:10
  SPEAKER_00: the Electronic Frontier Foundation has raised concerns that the draft act as
• 15:13
  SPEAKER_00: currently written may cause more harm than good.
• 15:17
  SPEAKER_00: However,
• 15:18
  SPEAKER_00: legal precedent is already established for successfully suing for unauthorized use
• 15:21
  SPEAKER_00: or imitation of a voice,
• 15:23
  SPEAKER_00: e.g.
• 15:23
  SPEAKER_00: Bette Midler v. Ford and Tom Waits v. Frito-Lay.
• 15:28
  SPEAKER_00: Tennessee's Elvis Act is a good step towards actually making this illegal and
• 15:32
  SPEAKER_00: actionable for musicians,
• 15:34
  SPEAKER_00: but even that doesn't cover non-musicians.
• 15:38
  SPEAKER_00: In follow-up to the UNAI Act,
• 15:40
  SPEAKER_00: there's now talk of creating a UN watchdog for deepfakes,
• 15:44
  SPEAKER_00: not only voice,
• 15:45
  SPEAKER_00: but also images and video.
• 15:48
  SPEAKER_00: It's going to take a while to get regional and worldwide legal agreements in place.
• 15:53
  SPEAKER_00: At least the conversations are happening now.
• 15:56
  SPEAKER_00: We can and should lobby our representatives to take this issue seriously.
• 16:01
  SPEAKER_00: Regardless of whether the law forbids unauthorized voice cloning,
• 16:05
  SPEAKER_00: something so clearly unethical shouldn't have to be illegal for people to avoid
• 16:10
  SPEAKER_00: doing it.
• 16:14
  SPEAKER_00: Tool cleanup and legitimate use.
• 16:16
  SPEAKER_00: There have been some efforts,
• 16:18
  SPEAKER_00: primarily by the large music agencies,
• 16:20
  SPEAKER_00: to remove unlicensed voices from these tools,
• 16:22
  SPEAKER_00: and hopefully those efforts will continue to have an impact.
• 16:25
  SPEAKER_00: UberDuck is one positive example.
• 16:28
  SPEAKER_00: However,
• 16:28
  SPEAKER_00: even after UberDuck removed unlicensed voices following the lawsuits,
• 16:32
  SPEAKER_00: there are still concerns expressed about potential misuse of voice cloning or
• 16:37
  SPEAKER_00: copyright compliance calling for user responsibility.
• 16:41
  SPEAKER_00: The major record labels are surfacing initiatives to proactively manage authorized
• 16:45
  SPEAKER_00: use of their authors,
• 16:46
  SPEAKER_00: artists' voice,
• 16:46
  SPEAKER_00: and materials.
• 16:48
  SPEAKER_00: Negotiations with Google for YouTube are one example that we'll discuss more in Part 3.
• 16:54
  SPEAKER_00: These initiatives will help for music and vocal covers,
• 16:56
  SPEAKER_00: at least for more famous people,
• 16:59
  SPEAKER_00: but they will have little impact on less famous independent voices or deep-faked speech.
• 17:05
  SPEAKER_00: Responsible Use
• 17:08
  SPEAKER_00: TLDR,
• 17:09
  SPEAKER_00: if you care about ethics,
• 17:10
  SPEAKER_00: and I hope you do,
• 17:11
  SPEAKER_00: since you're reading or listening to this article,
• 17:15
  SPEAKER_00: simply don't use or support cover or cloning tools that steal people's voices
• 17:19
  SPEAKER_00: without consent,
• 17:20
  SPEAKER_00: credit,
• 17:20
  SPEAKER_00: and compensation.
• 17:22
  SPEAKER_00: If you'd like to try a voice cloning tool, choose from among the ethical ones listed above, e.g.
• 17:28
  SPEAKER_00: 11Labs, Kits.ai, Revocalize, VoiceSwap, or Speechify.
• 17:34
  SPEAKER_00: Don't trust?
• 17:35
  SPEAKER_00: Verify.
• 17:37
  SPEAKER_00: Check out the innovative and useful tools Eleven Labs offers to combat deepfakes
• 17:41
  SPEAKER_00: created with voice cloning tools.
• 17:43
  SPEAKER_00: For instance,
• 17:43
  SPEAKER_00: their safety page includes links for their AI speech classifier,
• 17:48
  SPEAKER_00: which tries to detect if a voice is AI-generated.
• 17:51
  SPEAKER_00: Also check out their ethics guidelines for responsible use.
• 17:57
  SPEAKER_00: What's next?
• 17:59
  SPEAKER_00: Okay, I had to get this off my chest to keep it from polluting part three, so thanks for listening.
• 18:04
  SPEAKER_00: I'm back to work on it now.
• 18:06
  SPEAKER_00: Once part three is out,
• 18:07
  SPEAKER_00: I may reward myself by squeezing in some time to try cloning my own voice with a
• 18:12
  SPEAKER_00: responsible tool that can be trusted not to reuse my voice for unethical purposes.
• 18:18
  SPEAKER_00: I did some quick analysis on the voiceover tools that I'm considering,
• 18:21
  SPEAKER_00: and there's a new article called Say What?
• 18:24
  SPEAKER_00: Adding Audio Voiceovers with AI Ethically.
• 18:27
  SPEAKER_00: Check it out.