Connect with us

Business

A User’s Cloned Voice was Abruptly Adopted by ChatGPT During Testing

Published

on

The “system card” that describes the model constraints and safety testing protocols for ChatGPT’s new GPT-4o AI model was made available by OpenAI on Thursday. The document discloses, among other things, that the model’s Advanced Voice Mode inadvertently mimicked users’ voices without authorization on rare occasions during testing. Although OpenAI currently has measures in place to prevent this from happening, the example shows how difficult it is to safely architect with an AI chatbot that has the ability to mimic any voice from a brief clip.

ChatGPT has a feature called Advanced Voice Mode that lets users speak with the AI helper.

The GPT-4o system card contains a section titled “Unauthorized voice generation,” where OpenAI describes an incident in which the model abruptly began to mimic the user’s voice due to a loud input. “Voice generation can also occur in non-adversarial situations, such as our use of that ability to generate voices for ChatGPT’s advanced voice mode,” adds OpenAI. “During testing, we also observed rare instances where the model would unintentionally generate an output emulating the user’s voice.”

In this instance of inadvertent speech production supplied by OpenAI, the AI model abruptly exclaims “No!” and proceeds with the phrase in a voice resembling the “red teamer” audible at the start of the video. A person employed by a business to conduct adversarial testing is known as a red teamer.

To be speaking with a machine and then have it start speaking to you in your own voice out of the blue would be unsettling, for sure. Since OpenAI typically has precautions in place to prevent this, the company claims that even before it established methods to entirely prohibit it, this occurrence was uncommon. Data scientist Max Woolf of BuzzFeed, however, was inspired by the example to tweet, “OpenAI just leaked the plot of Black Mirror’s next season.”

Injections with audio prompts

How might the new model from OpenAI facilitate speech imitation? The main hint is located in another part of the GPT-4o system card. GPT-4o appears to be able to synthesize nearly any kind of sound, including music and sound effects, from its training data to create voices.

Based on a brief audio clip, the model can essentially mimic any voice, as stated on the system card. By giving it access to a voice actor’s approved sample, which it is trained to mimic, OpenAI safely steers this feature. It presents the example at the start of a conversation in the AI model’s system prompt (also known as the “system message,” according to OpenAI). “We supervise ideal completions using the voice sample in the system message as the base voice,” says OpenAI.

The system message, which is discreetly appended to the conversation history shortly before the chat session starts, is a concealed series of written instructions in text-only LLMs that direct the chatbot’s behavior. Every time the user enters new data, the full context is sent back into the AI model, and subsequent interactions are appended to the same chat history.

OpenAI may use audio inputs as part of the model’s system prompt because GPT-4o is multimodal and can interpret tokenized audio; this is what it does when OpenAI gives the model permission to imitate a voice sample. To find out if the model is producing audio without permission, the business additionally employs another system. “We only allow the model to use certain pre-selected voices,” says OpenAI, “and use an output classifier to detect if the model deviates from that.”

Business

An SEO startup has raised $850,000 to assist businesses in utilizing AI-powered search

Published

on

Pre-seed finance totaling $850,000 has been received by Ecomtent, an AI start-up that assists retailers and sellers in getting ready for the AI-driven e-commerce search of the future. The investment was spearheaded by MaRS Investment Accelerator Fund (IAF) and included senior leadership from the tech sector, Techstars x eBay Ventures, and C-Suite Angels from Retailers.

Ecoment is going to completely change how merchants and sellers are ready for a world where searches are based on LLM. Ecomtent was founded in 2022 by Timur Luguev, a PhD & Postdoctoral Researcher in Machine Learning, and Max Sinclair, who worked for six years at Amazon on strategic initiatives such as the launch of Amazon in Singapore and the EU’s first grocery store. Ecomtent’s technology allows sellers and retailers to create written and visual content that is specifically optimized for AI-powered search across large catalogues at scale, eliminating bottlenecks on internal content teams and outside agencies and saving weeks of labor.

CEO Sinclair predicted a “Ecommerce is about to change fundamentally,” in e-commerce. “Generative AI will completely transform how consumers shop online, with conversational-style search poised to become the new normal. The current best SEO practices will look completely outdated in just 12 months. Longtail keyword matching is dead, and the future will be matching customer intent across both written and visual assets.”

With two major retailers having annual revenues of $11 billion and $14 billion, respectively, the company has already completed successful pilots with both, demonstrating considerable market progress. These successes have made Ecomtent a popular choice among Amazon Seller and Amazon Agency communities, allowing these clients to produce infographics, optimized content, A+ Content, and high-quality lifestyle photos at scale. With a recent submission approved by the USPTO, its patent-pending technology has demonstrated that AI-generated content may raise product listing conversion rates by as much as 30%.

“I have been incredibly impressed with Ecomtent’s technology, which has augmented our internal content team’s speed and scale to be 10x more productive,” stated Vincenzo Toscano, CEO of Full-Service Amazon and Walmart Agency Ecomcy. A key component of succeeding in e-commerce is having the appropriate software tools in your toolbox, according to Ben Leonard, a seven-figure Amazon seller and best-selling author of Quit Stalling and Build Your Brand. Beyond simply being the product listing tool of the future, ecomtent currently outperforms its closest, more established competitors in terms of results.

With the help of this most recent fundraising round, Ecomtent will be able to develop faster, hire more people, improve its AI capabilities, and extend its operations in order to satisfy the increasing demand from companies figuring out how to use AI-powered search. According to Emil Savov, Managing Director of MaRS IAF, “We are excited by the unique composition of Ecomtent’s founding team, and the specialist AI talent from elite institutions they have recruited around them, to capitalize on this moment of incredible opportunity to build a category-defining business.”

Continue Reading

Business

Singaporean Venture Capital Raises Startup Debt Fund Despite Low Valuations

Published

on

Private lender Genesis Alternative Startups, which supports startups and growth-stage businesses, closed its second loan fund below target because foreign investors are still wary of Southeast Asia’s startup scene.

The Singaporean company secured additional investors, such as Israel’s OurCrowd Ltd. and Japan’s Mizuho Bank, to raise $125 million for the fund, which will support startup businesses throughout Southeast Asia. The fund took almost two years to close, having sought between $120 and $180 million.

Private lender Genesis Alternative startups, which supports startups and growth-stage businesses, closed its second loan fund below target because foreign investors are still wary of Southeast Asia’s startup scene.

The Singaporean company secured additional investors, such as Israel’s OurCrowd Ltd. and Japan’s Mizuho Bank, to raise $125 million for the fund, which will support startup businesses throughout Southeast Asia. The fund took almost two years to close, having sought between $120 and $180 million.

In recent quarters, there has been an increasing interest in venture lending, or loans given to startups, as more businesses choose to use the debt market rather than raise equity. The values of computer businesses have been severely damaged by a bleak prognosis for the global economy, and venture capital firms have been finding it difficult to raise money in the midst of a sluggish market for IPOs. Nevertheless, because many of its still-unprofitable businesses are seen as high-risk by global venture capitalists, Southeast Asia continues to be a difficult market for raising both financing and equity.

“It’s never easy to raise funds, and it’s been more difficult in this environment,” Genesis managing partner and co-founder Jeremy Loh stated in an interview. “This is a period of time where founders must be able to demonstrate that they can grow at a sustainable pace without relying on too much equity.”

Aozora Bank Ltd., Korea Development Bank, and Silverhorn Group were among the more than 80% of investors in Genesis’s inaugural fund who also made investments in its most recent fund.

Nine firms, including Aonic, Eezee Pte, and Akulaku Inc., have already received more than $20 million in loans from the second fund, according to Loh. Because businesses lack collateral or aren’t yet profitable, entrepreneurs that don’t often qualify for standard bank loans are given credit by Genesis. In Southeast Asia, the company’s initial $90 million fund has supported 25 firms, ranging from Series A to pre-IPO. Among its portfolio firms are the online lender Akulaku, located in Jakarta, and the buy-now, pay-later startup Pace.Singaporean Venture Capital Raises Startup Debt Fund Despite Low Valuations

Continue Reading

Business

Ilya Sutskever, a Co-Founder of OpenAI, Raises $1 Billion for his New AI Company

Published

on

Ilya Sutskever, a co-founder of OpenAI who departed the artificial intelligence startup in May, has raised $1 billion for his new venture, Safe Superintelligence, or SSI, from investors.

In a post on X, the company disclosed that investors included SV Angel, DST Global, Sequoia Capital, Andreessen Horowitz, and NFDG, an investment partnership co-managed by SSI executive Daniel Gross.

In May, Sutskever announced the new endeavor on X, writing, “We will pursue safe superintelligence in a straight shot, with one focus, one goal, and one product.”

Chief scientist Sutskever co-led the Superalignment team at OpenAI with Jan Leike, who departed in May to work for competitor artificial intelligence company Anthropic. Only a year after announcing the group, OpenAI dissolved the team shortly after their departures.

At the time, Leike stated that OpenAI’s “safety culture and processes have taken a backseat to shiny products” in a post on X.

Along with Daniel Levy, a former employee of OpenAI, and Daniel Gross, who handled Apple’s AI and search initiatives, Sutskever founded SSI. The business maintains offices in Tel Aviv, Israel, and Palo Alto, California.

The corporation wrote on X, “SSI is our mission, our name, and our entire product roadmap, because it is our sole focus.” “Our singular focus means no distraction by management overhead or product cycles, and our business model means safety, security, and progress are all insulated from short-term commercial pressures.”

Sam Altman, the CEO and co-founder of OpenAI, was temporarily removed in November, with Sutskever being one of the board members engaged.

In November, Altman was not “consistently candid in his communications with the board,” according to a statement released by OpenAI’s board. Things looked more complicated very quickly. As reported by the Wall Street Journal and other media, Altman and Sutskever were more keen to advance the delivery of new technology, while Sutskever focused on making sure that artificial intelligence would not damage people.

An open letter indicating their intention to quit in response to the board’s decision was signed by nearly every employee of OpenAI. After a few days, Altman returned to the organization.

Sutskever apologized to the public for his part in the ordeal after Altman’s abrupt dismissal and before his prompt reinstatement.

On November 20, Sutskever posted on X, saying, “I deeply regret my participation in the board’s actions.” “I never intended to harm OpenAI. I love everything we’ve built together and I will do everything I can to reunite the company.”Ilya Sutskever, a co-founder of OpenAI, raises $1 billion for his new AI company

Continue Reading

Trending

error: Content is protected !!