Search This Blog

Friday, April 28, 2023

Will AI Clones Replace Us?



 

Can AI Joanna fool the world and take a day off for the real Joanna?

Source: The Wall Street Journal
Date: April 28, 2023

Greg's Words

I remember the time I sold a PC, WordPerfect, and an HP Laser Series II to a legal assistant in a one-man lawyer office, in Grand Blanc, Michigan.  On her desk was an IBM Selectric.  She was proud of her ability to type four to five custom letters in 45 minutes; she was VERY good at her job and ran the office with razor-sharp efficiency.

Cutting edge, she convinced her boss to purchase a new contraption(IBM PS/2, AT?) to help her do more.

Her eyes lit up as letter after letter crept out of the LaserJet - blazing eight pages per minute.  She was giddy.

Suddenly, like a cloud drifting over the blueberry fields of Imlay City, her smile dissolved into shadow and I swear the blood rushed out of her cheeks.  Her cheer was supplanted by Doom.  She suddenly saw the whole of the Moon; her job, lifestyle, and future were shockingly in question. Obsolete.

Reading the Wall Street Journal entry by Joanna Stern, brought me back to that rainy, Mid-Michigan day in 1989.   Technology is so cool, it's Cold.

"AI Joanna" is the HP Series II of the day, a creation utilizing readily available resources producing less-than-perfect results.  Time is compressing, from 'normal' to 'AI Time', years are months, days hours and although AI Joanna is not 'perfect', somewhere in the wild, today, there is a perfect Replicant giving a local newscast or weather report.

We've put together a quick summary of the article.

Enjoy!

Key highlights:
  • AI-generated avatars and voice clones are becoming increasingly convincing, but their potential for misuse raises concerns.
  • The article focuses on AI Joanna, an AI-generated avatar and voice clone created using Synthesia and ElevenLabs.
  • Initiatives like the Adobe-led Content Authenticity Initiative are working to address the challenges associated with AI-generated content.
_________


Joanna Stern from The Wall Street Journal wrote, "Maybe you’ve played around with chatbots like OpenAI’s ChatGPT and Google’s Bard, or image generators like Dall-E. If you thought they blurred the line between AI and human intelligence, you ain’t seen—or heard—nothing yet." As AI tools continue to evolve, it's essential for users to acknowledge the two-sided nature they embody. While AI-generated avatars and voice clones offer convenience and efficiency, they also bring about inherent risks and challenges that need to be tackled. 
__________

The rapid advancements in AI-generated avatars and voice clones have led to tools like AI Joanna, which blur the line between human-generated and AI-generated content. AI Joanna is an AI-generated avatar created using Synthesia, a tool that produces deepfakes from recorded video and audio, and her voice clone was developed by ElevenLabs, an AI speech software developer. While AI Joanna showcases the potential of AI to improve productivity and reduce the drudgery of certain tasks, there are concerns about the potential for misuse and deception.
"I cued AI Joanna up with several things I knew Chase would ask, then dialed customer service. At the biometric step, when the automated system asked for my name and address, AI Joanna responded. Hearing my bot’s voice, the system recognized it as me and immediately connected to a representative. When our video intern called and did his best Joanna impression, the automated system asked for further verification." - J. Stern, WSJ
AI Joanna was designed to mimic the appearance and voice of Joanna Stern, a journalist. The avatar was created using around 30 minutes of video and nearly two hours of audio, which Synthesia used to train the AI model. AI Joanna can convincingly replicate Joanna's appearance and voice in short sentences, but her AI-generated nature becomes more apparent in longer text. While the technology has limitations in terms of natural expressions, gestures, and emotions, it is expected to improve over time.

The potential for misuse of AI-generated avatars and voice clones is significant. In one instance, AI Joanna's voice was able to fool Chase's credit card's voice biometric system, highlighting the risks associated with AI-generated voices in security and fraud prevention. The Federal Trade Commission is already warning about AI-voice-related scams and companies like ElevenLabs and Synthesia is working on measures to ensure the responsible use of their technology.

One of the challenges in addressing AI-generated content is the difficulty of spotting synthetic audio and video. Hany Farid, a digital forensics expert at the University of California, Berkeley, points out that it is difficult to train everyone to be an AI detective. While AI-generated content like AI Joanna is easy to spot for now, the technology will only get better, making it harder to distinguish between real and artificial personas.

Initiatives like the Adobe-led Content Authenticity Initiative are working to address the challenges associated with AI-generated content. The initiative, which has over 1,000 members, including media and tech companies, academics, and more, aims to create an embedded "nutrition label" for media content. This would allow photos, videos, and audio on the internet to come with verifiable information attached, making it easier to authenticate the source and nature of the content. Synthesia, the company behind AI Joanna's avatar, is a member of the initiative.

In the meantime, AI-generated content like AI Joanna serves as a reminder of the advancements and challenges that AI tools present. While these technologies have the potential to improve productivity and allow humans to focus on other tasks, they also raise concerns about fraud, scams, and the need for authenticating content. It is essential for individuals, businesses, and regulators to be aware of the potential implications of AI-generated content and to stay informed about the latest developments in this rapidly evolving field.

As AI tools continue to evolve, it's beneficial for users to acknowledge the two-sided nature they embody. While AI-generated avatars and voice clones offer convenience and efficiency, they also bring about inherent risks and challenges that need to be tackled. 

As the ominous shadow of AI tools stretches further, users are beckoned to confront the duality that lies within. These AI-generated avatars and voice clones weave a tapestry of convenience and efficiency, but hidden beneath lurks the perilous risks and challenges they harbor. Standing vigilant and wielding AI technologies responsibly, we can journey through the realm of innovation, seizing its blessings while warding off the lurking darkness.

 ___________
References:

"Testing AI Joanna: A Personal Deepfake Avatar," The Wall Street Journal, April 28, 2023

Tweet: "Can AI Joanna fool the world and take a day off for the real Joanna? Find out in our latest article on #AI #deepfakes and the challenges they present. #Synthesia #ElevenLabs"

LinkedIn Intro: Explore the world of AI-generated avatars and voice clones as we dive into the experiment of AI Joanna, a deepfake avatar that attempts to replace the real Joanna Stern in videos, meetings, and phone calls. Discover the advancements, limitations, and concerns surrounding this technology.

Keywords: AI Joanna, Synthesia, deep fake avatar, ElevenLabs, AI voice clone, biometric system, Content Authenticity Initiative

Image prompt: A side-by-side comparison of the real Joanna Stern and AI Joanna avatar

Search question: How effective are AI-generated avatars and voice clones in mimicking human behavior?

Real song that matches the theme: "Fake You Out" by Twenty-One Pilots matches the theme of the article because it metaphorically relates to the concept of AI-generated avatars and voice clones attempting to "fake out" or deceive people by mimicking human appearance and behavior. The song's title and lyrics can be interpreted as a commentary on the challenge of distinguishing between real and artificial personas, which is relevant to the topic of deep fakes and AI-generated content.

In summary, AI-generated avatars and voice clones like AI Joanna are becoming increasingly sophisticated, blurring the line between human-generated and AI-generated content. While there are still limitations in terms of natural expressions, gestures, and emotions, these technologies are expected to improve over time.

As AI tools become more advanced, they may provide humans with more time to focus on other tasks. However, the potential for misuse of AI-generated voice and video is significant, raising concerns about fraud, scams, and the need for authenticating content. Initiatives like the Adobe-led Content Authenticity Initiative are working to address these concerns by creating verifiable information for media content.

In the meantime, AI-generated content like AI Joanna serves as a reminder of the advancements and challenges that these technologies present. It is essential for individuals, businesses, and regulators to be aware of the potential implications of AI-generated content and to stay informed about the latest developments in this rapidly evolving field.
________


No comments:

Post a Comment

Contact Me

Greg Walters, Incorporated
greg@grwalters.com
262.370.4193