AI Replicates My Voice: Experimenting with ElevenLabs and the Future of AI Regulation
As the world of AI continues to evolve, so does the potential for these models to replicate not only tasks but human elements like voice. This concept piqued my interest recently after watching Oprah’s special AI and the Future of Us, featuring the ever-popular tech reviewer Marques Brownlee. He demonstrated how ElevenLabs, an online tool, could replicate voices with startling accuracy using just a 30-minute audio sample.
Inspired by this, I thought: why not try it myself? Why not take this opportunity to merge two important areas of my life—my work advocating for ethical AI and my personal voice—to see how well AI could replicate me? So, I set off on a journey to train the AI using a recording of me reading some of the key articles from my nonprofit, Voice for Change Foundation, particularly around Senate Bill 1047. This bill, currently awaiting the signature of California Governor Gavin Newsom, represents a crucial turning point in regulating AI, protecting jobs, and ensuring AI's development is accountable and ethical.
The Experiment Begins: Recording in My Car
Since the tool requires a relatively quiet environment, and I spend a lot of time in my car—an unintended sanctuary of reflection and productivity—it became the ideal spot for this experiment. With my phone in hand, I began recording, choosing to read aloud an open letter I had written to Governor Newsom on Labor Day. This letter was part of my advocacy work, urging the governor to sign Senate Bill 1047 for the safeguarding of not just human jobs but human dignity in the face of rapidly growing AI technologies.
I admit, I felt a bit apprehensive about hearing what the AI version of my voice would sound like. Would it capture my accent, tone, or the emotional nuances I express when discussing topics close to my heart? I knew it wouldn’t be perfect, but technology has advanced so far, so I was ready to be both amazed and unsettled by what it would produce.
The Power of Voice Replication
For 30 minutes, I read various articles and letters from the Voice for Change website, particularly those focusing on SB 1047 and its potential impact on workforce preservation and AI regulation. This bill, led by Senator Scott Wiener, calls for necessary safeguards that would hold AI companies accountable for the potential harms their models might cause. It also requires large-scale AI developers to implement safety plans and whistleblower protections to mitigate risks to the economy and society.
Reading my own words to train this AI, I couldn't help but think of the irony—using AI to generate my voice for a cause centered on regulating AI itself. It felt both exciting and somewhat dystopian. The very technology I’m advocating to regulate is now being used to simulate me.
Why SB 1047 Matters
For those unfamiliar, Senate Bill 1047 is more than just another piece of legislation; it’s a safeguard for the future. AI’s potential to disrupt industries is well-documented, and without regulatory frameworks in place, we’re venturing into dangerous territory. As I read my letter to the governor aloud for this AI experiment, the urgency of this moment struck me once more.
The letter highlighted the risks of allowing AI to develop unchecked, where it could influence elections, disrupt industries, and possibly even be weaponized. These aren’t hypothetical fears but real, documented possibilities. My letter, which was also generated in part with the assistance of AI (yes, irony again), aims to persuade Governor Newsom to sign the bill, emphasizing that the stakes are too high for inaction.
Reflection: What’s at Stake?
What this experiment truly brought to light is the delicate balance we must maintain between embracing innovation and ensuring that innovation does not come at the cost of humanity. Training an AI to sound like me was an exciting exercise in pushing the boundaries of what’s possible. However, it also served as a reminder of how rapidly we are advancing into a world where human identity, work, and voice can be replicated, commodified, or even exploited if left unchecked.
While I wait to hear the results of this voice replication (which, according to ElevenLabs, takes four to six hours), I’m reminded once again of the importance of regulation. Senate Bill 1047 isn't just about protecting jobs; it’s about protecting the integrity of human experience and ensuring that AI serves us, not the other way around.
Conclusion: The Future of AI and Human Rights
As technology evolves, so must our laws. The experiment with ElevenLabs was a small reminder of how powerful AI can be and why we need Senate Bill 1047 signed into law. It’s not about restricting innovation—it's about ensuring that innovation remains ethical, that AI companies are held accountable, and that our society is safeguarded from the possible consequences of unregulated AI growth.
Stay tuned as I explore more on how my AI-generated voice turns out, and in the meantime, let's continue to push for responsible AI governance. Because the future of AI isn’t just about what it can do—it’s about how we decide to shape its impact on the world.
