OpenAI announced Voice Engine, a text-to-speech AI model for creating synthetic voices based on a 15-second segment of recorded audio, but then, out of concerns for its misuse, the company decided to delay its introduction.
In a blog post explaining its decision, the company recommended changes it said society will have to make to prepare for the introduction of the technology. Among them: Phasing out voic -based authentication as a security measure for accessing bank accounts and other sensitive information.
Banks and financial services providers are among the first companies to be targeted. Researchers and reporters have shown that voice-cloning technology can be used to break into bank accounts that use voice authentication (such as Chase’s Voice ID), which prompted US senator Sherrod Brown of Ohio, the chair of the U.S. Senate Committee on Banking, Housing, and Urban Affairs, to send a letter to the CEOs of several major banks in May 2023 to inquire about the security measures banks are taking to counteract AI-powered risks.
Voice cloning tech in general is not particularly new—there have been several AI voice synthesis models since 2022, and the tech is active in the open source community with packages like OpenVoice and XTTSv2, notes Wired magazine. But the idea that OpenAI considered letting anyone use its particular brand of voice tech is notable. And in some ways, the company’s reticence to release it fully might be the bigger story.
Even if OpenAI never widely releases its Voice Engine, the ability to clone voices has already caused trouble in society through phone scams like when someone imitates a loved one’s voice or election campaign robocalls featuring cloned voices from politicians like U.S. President Joe Biden.
In a cyber threat report released this week Microsoft said Chinese influence campaigns are both amplifying AI-generated media that benefits their strategic narratives, as well as creating their own video, memes, and audio content. Such tactics have been used in campaigns stoking divisions within the United States and exacerbating rifts in the Asia-Pacific region—including Taiwan, Japan, and South Korea, says the report.
Western intelligence officials have said they have growing concerns about how AI tools could be used to flood elections this year with misleading videos or other content, including in the 2024 U.S. presidential contest. Security experts have said fake AI-generated audio clips pose an especially acute threat because they are relatively easy to manufacture and have been shown to dupe audiences easily.
In its blog post OpenAI says it recognizes that generating speech that resembles people’s voices has serious risks, especially in a year that will see high-stakes elections in over 50 countries. The U.S. AI research organization said the partners testing Voice Engine today have agreed to usage policies, which prohibit the impersonation of another individual or organization without consent or legal right. In addition, OpenAI’s terms with these partners require explicit and informed consent from the original speaker and we don’t allow developers to build ways for individual users to create their own voices. Partners must also clearly disclose to their audience that the voices they’re hearing are AI-generated. Finally, Open AI says it has implemented a set of safety measures, including watermarking to trace the origin of any audio generated by Voice Engine, as well as proactive monitoring of how it’s being used.
“We believe that any broad deployment of synthetic voice technology should be accompanied by voice authentication experiences that verify that the original speaker is knowingly adding their voice to the service and a no-go voice list that detects and prevents the creation of voices that are too similar to prominent figures,” says OpenAI’s blog post. “It’s important that people around the world understand where this technology is headed, whether we ultimately deploy it widely ourselves or not.”
IN OTHER NEWS THIS WEEK:
QUANTUM COMPUTING
Microsoft and Quantinuum Announce Quantum Computing Breakthrough
Microsoft and Quantinuum, a quantum computing company formed by the merger of Cambridge Quantum and Honeywell Quantum Solutions, announced a major breakthrough in quantum error correction. Using Quantinuum’s ion-trap hardware and Microsoft’s new qubit-virtualization system, the team was able to run more than 14,000 experiments without a single error. This new system also allowed the team to check the logical qubits and correct any errors it encountered without destroying the logical qubits. “This is a crucial milestone on our path to building a hybrid supercomputing system that can transform research and innovation across many industries,” Microsoft said in a blog posting.” It is made possible by the collective advancement of quantum hardware, qubit virtualization and correction, and hybrid applications that take advantage of the best of AI, supercomputing, and quantum capabilities. With a hybrid supercomputer powered by 100 reliable logical qubits, organizations would start to see scientific advantage, while scaling closer to 1,000 reliable logical qubits would unlock commercial advantage.”
FUSION
South Korean Research Center Announces Nuclear Fusion Breakthrough
Scientists in South Korea have announced a new world record for the length of time they sustained temperatures of 100 million degrees Celsius — seven times hotter than the sun’s core — during a nuclear fusion experiment, in what they say is an important step forward for this futuristic energy technology. Nuclear fusion seeks to replicate the reaction that make the sun and stars shine, by fusing together two atoms to unleash huge amounts of energy. Often referred to as the holy grail of climate solutions, fusion has the potential to provide limitless energy without planet-warming carbon pollution.
FINANCIAL SERVICES
JPMorgan Chase Launches Digital Media Business
JPMorgan Chase launched a new digital media business that would allow advertisers to target the Wall Street bank’s 80 million customers based on their spending data. The new platform, called Chase Media Solutions, will combine the scale of a retail media network with Chase-owned transaction data that will help brands to precisely target customers, the company said.
ARTIFICIAL INTELLIGENCE
Google Considers Charging For AI-Powered Search
Google is considering charging for new “premium” features powered by generative artificial intelligence, in what would be the biggest ever shake-up of its search business, reports The Financial Times. The proposed revamp to its cash cow search engine would mark the first time the company has put any of its core product behind a paywall, and shows it is still grappling with a technology that threatens its advertising business, almost a year and a half after the debut of ChatGPT.
U.S. and UK Sign Landmark Agreement On Testing Safety Of AI
The U.S. and UK have signed a landmark agreement on artificial intelligence, as the allies become the first countries to formally co-operate on how to test and assess risks from emerging AI models. The agreement, signed on April1 in Washington by UK science minister Michelle Donelan and US commerce secretary Gina Raimondo, lays out how the two governments will pool technical knowledge, information and talent on AI safety. The deal represents the first bilateral arrangement on AI safety in the world and comes as governments push for greater regulation of the existential risks from new technology, such as its use in damaging cyber attacks or designing bioweapons.
White House Orders U.S. Federal Agencies To Name Chief AI Officers
The White House is ordering all federal agencies to name chief artificial intelligence officers to oversee the federal government’s various approaches to AI and manage the risks that the rapidly evolving technologies might pose.
To access more of The Innovator’s News In Context articles click here.
