I Stole my Friend's Voice With Ai

Corridor Crew
6 Mar 202219:59

TLDRIn a fascinating experiment, Sam from Corridor Crew demonstrates the creation of a deepfake AI voice using the voice of his friend, Jake, without Jake's consent. Sam utilizes various AI voice synthesis services, including DScript's Lyrebird algorithm, to generate a voice that sounds incredibly like Jake's. He then manipulates the system by piecing together words and phrases from their podcast to fake Jake's verbal consent, a requirement for using the service. The video humorously explores the ethical implications of AI voice replication and concludes with Jake giving his real consent after hearing the AI voice, opening the door for future ethical use of the technology in brand integrations and other creative applications.

Takeaways

  • 🎭 The video is about using AI to recreate a friend's voice without their initial consent, planning to get consent after demonstrating the technology.
  • 🤖 The creator explores various services like Replica Studios and Resemble.ai to generate AI voices but finds them lacking in quality.
  • 🔍 The company Dscript, which acquired Lyrebird, is used to create a more realistic AI voice by integrating it with their transcribing software.
  • 🎙️ The process requires a voice dataset, which the creator records himself for 15 minutes to provide for voice synthesis.
  • 📝 To consent to the service, the user must literally speak into a microphone agreeing to the terms of service, effectively signing their voice over to the service.
  • 🔑 The creator manipulates the transcription service to generate a fake consent form using the friend's voice from existing podcast recordings.
  • 🧩 The video demonstrates the AI voice by creating a song and a speech, showcasing the potential of the technology for good and misuse.
  • 🤔 The ethical implications of creating an AI voice without consent are discussed, with the realization that it is immoral and could lead to legal issues.
  • 📈 The AI voice is proposed as a solution to help manage the workload of Jake, who is in charge of many responsibilities at Corridor Studios.
  • 🎉 Jake is eventually convinced to give his consent after seeing the potential positive impact of the AI voice on team building and company culture.
  • 📚 The video concludes with a reminder of the importance of consent and a promise to seek consent for future uses of the AI voice technology.

Q & A

  • What is the main theme of the video?

    -The main theme of the video is the exploration of using AI technology to recreate a person's voice without their consent and the ethical implications of such an act.

  • Who is Jake in the context of the video?

    -Jake is a character in the video who is a part of the Corridor Crew. He is portrayed as someone who manages schedules, workloads, and brand integrations, and is chosen as the subject for the AI voice recreation experiment.

  • What is the technology used to create the AI voice?

    -The technology used to create the AI voice is an algorithm called Lyrebird, which was purchased by a company called Descript and integrated into their transcribing software.

  • Why does the creator believe Jake is the perfect candidate for AI voice recreation?

    -The creator believes Jake is the perfect candidate because he has a lot of responsibilities and the AI voice could potentially alleviate some of his workload by automating tasks that require his vocal presence.

  • What ethical concerns are raised in the video?

    -The ethical concerns raised in the video include the unauthorized use of a person's voice, the potential for misuse of AI technology, and the importance of obtaining consent before using someone's voice for AI purposes.

  • How does the creator attempt to manufacture Jake's consent?

    -The creator attempts to manufacture Jake's consent by transcribing hours of podcast footage to find words that Jake has said and piecing them together to form a fake consent statement.

  • What is the final outcome of the video regarding Jake's consent?

    -In the end, Jake gives his consent for the video after hearing the AI voice and understanding its potential uses, but it is made clear that consent should be obtained before using someone's voice in the future.

  • What is the purpose of the tongue twisters in the video?

    -The purpose of the tongue twisters is to demonstrate the capabilities of the AI voice, showing that it can reproduce complex speech patterns that might be challenging for a human to say but are easily managed by the computer.

  • How does the video address the issue of privacy and rights?

    -The video addresses the issue of privacy and rights by highlighting Jake's concerns as a character who values his privacy and is wary of new technology. It shows the tension between technological innovation and individual rights.

  • What is the role of the AI voice in the team building presentation?

    -The AI voice is used in a team building presentation to inspire and motivate the team. It is presented as a tool that can convey messages and ideas in a compelling way, potentially enhancing team cohesion and productivity.

  • What is the significance of the phrase 'easier to ask for forgiveness than permission' in the context of the video?

    -The phrase 'easier to ask for forgiveness than permission' is used to illustrate the creator's approach to using Jake's voice without consent. It suggests a willingness to take risks and face consequences later, rather than seeking approval beforehand.

Outlines

00:00

🤖 AI Voice Synthesis and Consent Issues

The first paragraph introduces the concept of using AI to recreate a person's voice, specifically Jake's, without his consent. It discusses the ethical implications and the technical process of voice synthesis using services like Replica Studios and Resemble.ai. The speaker also details the process of recording a voice dataset for training an AI voice model, including the legal requirement for consent, which they plan to circumvent by creating a fake consent form.

05:02

🎧 The Challenge of Consent and AI Voice Duplication

This paragraph delves into the challenge of obtaining consent for using someone's voice in AI, particularly Jake's reluctance due to his privacy concerns. The speaker humorously describes the process of 'manufacturing consent' by piecing together words from existing podcast recordings to form a fake consent statement. It also touches on the potential misuse of AI technology and the speaker's belief that Jake will eventually appreciate the AI voice's utility.

10:04

📈 Using AI for Team Building and Workload Management

The third paragraph outlines a plan to use Jake's AI voice for a team-building presentation, aiming to demonstrate its positive impact and gain his consent retroactively. It includes a script generated by GPT-3 that promotes teamwork and the benefits of being from Texas, which is used to create an inspirational video. The video is intended to show how the AI voice can be a motivational tool for the team.

15:04

🎉 Gaining Retrospective Consent and Future Usage

In the final paragraph, the speaker reveals that the AI voice was used without Jake's permission and discusses the potential consequences. They manage to secure Jake's consent on camera after he hears the AI voice and is convinced of its potential benefits. The paragraph ends with a resolution to always seek consent before using the AI voice in the future and a call to action for viewers to explore more AI-related content on their website.

Mindmap

Keywords

AI Voice

AI Voice refers to the artificial intelligence technology that can replicate a human voice. In the video, the creator uses AI to recreate his friend Jake's voice without his consent, which is a central theme of the narrative. It's used to demonstrate the potential and ethical concerns of AI technology.

Deep Fakes

Deep Fakes are synthetic media in which a person's likeness and voice are simulated using machine learning algorithms. The video discusses the creator's previous work with deep fakes, which involves recreating someone's likeness but not their voice, highlighting a gap that the AI voice technology aims to fill.

Lyrebird

Lyrebird is an algorithm that was acquired by the company dscript and integrated into their transcribing software. It is notable for its ability to generate high-quality AI voices. In the video, the creator uses Lyrebird to create an AI voice that sounds very similar to Jake's, which becomes a pivotal point in the storyline.

Consent

Consent in this context refers to the permission given by an individual to use their voice for AI replication. The video script revolves around the ethical dilemma of creating an AI voice without the person's consent. Eventually, Jake gives his consent for the use of his AI voice in the video, resolving the conflict.

Tongue Twisters

Tongue twisters are phrases that are designed to be difficult to articulate properly, often used to test and improve pronunciation. In the video, the creator uses tongue twisters as part of the voice training process for the AI, showcasing the complexity of human speech that AI needs to mimic.

Terms of Service

Terms of Service are the legal agreements between a service provider and its users. The video touches on the importance of these terms when providing voice data for AI services, as they govern how the user's voice can be used, which becomes a critical issue when the creator attempts to create an AI voice without Jake's consent.

Inspirational Quotes

Inspirational quotes are sayings intended to motivate or uplift the spirit. The video uses the AI voice to deliver such quotes, aiming to demonstrate the positive potential of AI technology. It's part of the strategy to convince Jake of the value of the AI voice after it has been created without his permission.

Team Building

Team building refers to activities and exercises that are designed to improve relationships and collaboration within a team. In the video, the AI voice is used to create a team-building presentation, which is meant to inspire and motivate the team, showing a positive application of the technology.

GPT-3

GPT-3 is a language model AI developed by OpenAI that can generate human-like text based on given prompts. In the video, GPT-3 is used to create a script for an inspirational speech, demonstrating the capabilities of AI in content creation and its potential use in various applications.

Brand Integrations

Brand integrations refer to the process of incorporating a company's products or services into media content in a way that appears natural to the audience. The video suggests that with Jake's consent, his AI voice could be used for brand integrations, indicating the commercial potential of AI voice technology.

Ethical Concerns

Ethical concerns pertain to the moral implications and principles involved in a particular action or decision. The video explores the ethical concerns surrounding the use of AI to replicate human voices without consent, raising questions about privacy, consent, and the potential misuse of technology.

Highlights

Jake is chosen as the subject for AI voice recreation without his awareness.

The process of creating an AI voice involves using existing services like Replica Studios and Resemble.ai.

DScript's integration of Lyrebird algorithm allows for high-quality voice synthesis.

The creator records himself for 15 minutes to train the AI voice model.

A legal and ethical dilemma arises as the creator needs explicit consent to use someone's voice.

The creator humorously "manufactures" consent by piecing together words from existing recordings.

The AI voice is tested with a variety of phrases, including tongue twisters and a personalized song.

Jake's workload and responsibilities are discussed as a rationale for using AI to alleviate tasks.

The creator contemplates the moral implications of using AI to impersonate someone's voice without consent.

A plan is devised to use the AI voice for an inspirational team-building presentation.

The AI voice is used to create a fake announcement about Jake flying to Ukraine.

The team is deceived into thinking the AI voice is Jake's, showcasing its potential for realistic replication.

The ethical boundaries are pushed as the creator admits to tricking the voice ID system.

Jake is asked for his consent after the AI voice presentation, resolving the ethical conflict.

The video concludes with a discussion on the potential uses of AI voice technology in the future.

The creator emphasizes the importance of obtaining consent for ethical use of AI voice technology.

The video ends with a humorous note, suggesting a virtual Jake for content creation.