Toolbox

NVIDIA and WPP (NYSE: WPP) are developing a content engine that harnesses NVIDIA Omniverse™ and AI to enable creative teams to produce high-quality commercial content faster, more efficiently and at scale while staying fully aligned with a client’s brand.

The new engine connects an ecosystem of 3D design, manufacturing and creative supply chain tools, including those from Adobe and Getty Images, letting WPP’s artists and designers integrate 3D content creation with generative AI. This enables WPP’s clients to reach consumers in highly personalized and engaging ways, while preserving the quality, accuracy and fidelity of their company’s brand identity, products and logos.

NVIDIA founder and CEO Jensen Huang unveiled the engine in a demo during his COMPUTEX keynote address, illustrating how clients can work with teams at marketing services organization WPP to make large volumes of brand advertising content such as images or videos and experiences like 3D product configurators more tailored and immersive.

“The world’s industries, including the $700 billion digital advertising industry, are racing to realize the benefits of AI,” Huang said. “With Omniverse Cloud and generative AI tools, WPP is giving brands the ability to build and deploy product experiences and compelling content at a level of realism and scale never possible before.”

“Generative AI is changing the world of marketing at incredible speed,” said Mark Read, CEO of WPP. “Our partnership with NVIDIA gives WPP a unique competitive advantage through an AI solution that is available to clients nowhere else in the market today. This new technology will transform the way that brands create content for commercial use, and cements WPP’s position as the industry leader in the creative application of AI for the world’s top brands.”

An Engine for Creativity
The new content engine has at its foundation Omniverse Cloud — a platform for connecting 3D tools, and developing and operating industrial digitalization applications. This allows WPP to seamlessly connect its supply chain of product-design data from software such as Adobe’s Substance 3D tools for 3D and immersive content creation, plus computer-aided design tools to create brand-accurate, photoreal digital twins of client products.

WPP uses responsibly trained generative AI tools and content from partners such as Adobe and Getty Images so its designers can create varied, high-fidelity images from text prompts and bring them into scenes. This includes Adobe Firefly, a family of creative generative AI models, and exclusive visual content from Getty Images created using NVIDIA Picasso, a foundry for custom generative AI models for visual design.

With the final scenes, creative teams can render large volumes of brand-accurate, 2D images and videos for classic advertising, or publish interactive 3D product configurators to NVIDIA Graphics Delivery Network, a worldwide, graphics streaming network, for consumers to experience on any web device.

In addition to speed and efficiency, the new engine outperforms current methods, which require creatives to manually create hundreds of thousands of pieces of content using disparate data coming from disconnected tools and systems.

The partnership with NVIDIA builds on WPP’s existing leadership position in emerging technologies and generative AI, with award-winning campaigns for major clients around the world.
The new content engine will soon be available exclusively to WPP’s clients around the world.

The White House on Tuesday announced new efforts to guide federally backed research on artificial intelligence as the Biden administration looks to get a firmer grip on understanding the risks and opportunities of the rapidly evolving technology.

Among the moves unveiled by the administration was a tweak to the United States' strategic plan on artificial intelligence research, which was last updated in 2019, to add greater emphasis on international collaboration with allies.

White House officials on Tuesday were also hosting a listening session with workers on their firsthand experiences with employers' use of automated technologies for surveillance, monitoring, evaluation, and management. And the U.S. Department of Education's Office of Educational Technology issued a report focused on the risks and opportunities related to AI in education.

"The report recognizes that AI can enable new forms of interaction between educators and students, help educators address variability in learning, increase feedback loops, and support educators," the White House said in a statement. "It also underscores the risks associated with AI — including algorithmic bias — and the importance of trust, safety, and appropriate guardrails."

The U.S. government and private sector in recent months have begun more publicly weighing the possibilities and perils of artificial intelligence.

Tools like the popular AI chatbot ChatGPT have sparked a surge of commercial investment in other AI tools that can write convincingly human-like text and churn out new images, music and computer code. The ease with which AI technology can be used to mimic humans has also propelled governments around the world to consider how it could take away jobs, trick people and spread disinformation.

Last week, Senate Majority Leader Chuck Schumer said Congress "must move quickly" to regulate artificial intelligence. He has also convened a bipartisan group of senators to work on legislation.

The latest efforts by the administration come after Vice President Kamala Harris met earlier this month with the heads of Google, Microsoft, ChatGPT-creator OpenAI and Anthropic. The administration also previously announced an investment of $140 million to establish seven new AI research institutes.

The White House Office of Science and Technology Policy on Tuesday also issued a new request for public input on national priorities "for mitigating AI risks, protecting individuals' rights and safety, and harnessing AI to improve lives."

The head of the artificial intelligence company that makes ChatGPT told Congress on Tuesday that government intervention will be critical to mitigating the risks of increasingly powerful AI systems.

"As this technology advances, we understand that people are anxious about how it could change the way we live. We are too," OpenAI CEO Sam Altman said at a Senate hearing.

Altman proposed the formation of a U.S. or global agency that would license the most powerful AI systems and have the authority to "take that license away and ensure compliance with safety standards."

His San Francisco-based startup rocketed to public attention after it released ChatGPT late last year. The free chatbot tool answers questions with convincingly human-like responses.

What started out as a panic among educators about ChatGPT's use to cheat on homework assignments has expanded to broader concerns about the ability of the latest crop of "generative AI" tools to mislead people, spread falsehoods, violate copyright protections and upend some jobs.

And while there's no immediate sign Congress will craft sweeping new AI rules, as European lawmakers are doing, the societal concerns brought Altman and other tech CEOs to the White House earlier this month and have led U.S. agencies to promise to crack down on harmful AI products that break existing civil rights and consumer protection laws.

Sen. Richard Blumenthal, the Connecticut Democrat who chairs the Senate Judiciary Committee's subcommittee on privacy, technology and the law, opened the hearing with a recorded speech that sounded like the senator, but was actually a voice clone trained on Blumenthal's floor speeches and reciting ChatGPT-written opening remarks.

The result was impressive, said Blumenthal, but he added, "What if I had asked it, and what if it had provided, an endorsement of Ukraine surrendering or (Russian President) Vladimir Putin's leadership?"

The overall tone of senators' questioning was polite Tuesday, a contrast to past congressional hearings in which tech and social media executives faced tough grillings over the industry's failures to manage data privacy or counter harmful misinformation. In part, that was because both Democrats and Republicans said they were interested in seeking Altman's expertise on averting problems that haven't yet occurred.

Blumenthal said AI companies ought to be required to test their systems and disclose known risks before releasing them, and expressed particular concern about how future AI systems could destabilize the job market. Altman was largely in agreement, though had a more optimistic take on the future of work.

Pressed on his own worst fear about AI, Altman mostly avoided specifics, except to say that the industry could cause "significant harm to the world" and that "if this technology goes wrong, it can go quite wrong."

But he later proposed that a new regulatory agency should impose safeguards that would block AI models that could "self-replicate and self-exfiltrate into the wild" — hinting at futuristic concerns about advanced AI systems that could manipulate humans into ceding control.

That focus on a far-off "science fiction trope" of super-powerful AI could make it harder to take action against already existing harms that require regulators to dig deep on data transparency, discriminatory behavior and potential for trickery and disinformation, said a former Biden administration official who co-authored its plan for an AI bill of rights.

"It's the fear of these (super-powerful) systems and our lack of understanding of them that is making everyone have a collective freak-out," said Suresh Venkatasubramanian, a Brown University computer scientist who was assistant director for science and justice at the White House Office of Science and Technology Policy. "This fear, which is very unfounded, is a distraction from all the concerns we're dealing with right now."

OpenAI has expressed those existential concerns since its inception. Co-founded by Altman in 2015 with backing from tech billionaire Elon Musk, the startup has evolved from a nonprofit research lab with a safety-focused mission into a business. Its other popular AI products include the image-maker DALL-E. Microsoft has invested billions of dollars into the startup and has integrated its technology into its own products, including its search engine Bing.

Altman is also planning to embark on a worldwide tour this month to national capitals and major cities across six continents to talk about the technology with policymakers and the public. On the eve of his Senate testimony, he dined with dozens of U.S. lawmakers, several of whom told CNBC they were impressed by his comments.

Also testifying were IBM's chief privacy and trust officer, Christina Montgomery, and Gary Marcus, a professor emeritus at New York University who was among a group of AI experts who called on OpenAI and other tech firms to pause their development of more powerful AI models for six months to give society more time to consider the risks. The letter was a response to the March release of OpenAI's latest model, GPT-4, described as more powerful than ChatGPT.

The panel's ranking Republican, Sen. Josh Hawley of Missouri, said the technology has big implications for elections, jobs and national security. He said Tuesday's hearing marked "a critical first step towards understanding what Congress should do."

A number of tech industry leaders have said they welcome some form of AI oversight but have cautioned against what they see as overly heavy-handed rules. Altman and Marcus both called for an AI-focused regulator, preferably an international one, with Altman citing the precedent of the U.N.'s nuclear agency and Marcus comparing it to the U.S. Food and Drug Administration. But IBM's Montgomery instead asked Congress to take a "precision regulation" approach.

"We think that AI should be regulated at the point of risk, essentially," Montgomery said, by establishing rules that govern the deployment of specific uses of AI rather than the technology itself.

Matt O'Brien is an AP technology writer
 

No sooner did ChatGPT get unleashed than hackers started "jailbreaking" the artificial intelligence chatbot — trying to override its safeguards so it could blurt out something unhinged or obscene.

But now its maker, OpenAI, and other major AI providers such as Google and Microsoft, are coordinating with the Biden administration to let thousands of hackers take a shot at testing the limits of their technology.

Some of the things they'll be looking to find: How can chatbots be manipulated to cause harm? Will they share the private information we confide in them to other users? And why do they assume a doctor is a man and a nurse is a woman?

"This is why we need thousands of people," said Rumman Chowdhury, lead coordinator of the mass hacking event planned for this summer's DEF CON hacker convention in Las Vegas that's expected to draw several thousand people. "We need a lot of people with a wide range of lived experiences, subject matter expertise and backgrounds hacking at these models and trying to find problems that can then go be fixed."

Anyone who's tried ChatGPT, Microsoft's Bing chatbot or Google's Bard will have quickly learned that they have a tendency to fabricate information and confidently present it as fact. These systems, built on what's known as large language models, also emulate the cultural biases they've learned from being trained upon huge troves of what people have written online.

The idea of a mass hack caught the attention of U.S. government officials in March at the South by Southwest festival in Austin, Texas, where Sven Cattell, founder of DEF CON's long-running AI Village, and Austin Carson, president of responsible AI nonprofit SeedAI, helped lead a workshop inviting community college students to hack an AI model.

Carson said those conversations eventually blossomed into a proposal to test AI language models following the guidelines of the White House's Blueprint for an AI Bill of Rights — a set of principles to limit the impacts of algorithmic bias, give users control over their data and ensure that automated systems are used safely and transparently.

There's already a community of users trying their best to trick chatbots and highlight their flaws. Some are official "red teams" authorized by the companies to "prompt attack" the AI models to discover their vulnerabilities. Many others are hobbyists showing off humorous or disturbing outputs on social media until they get banned for violating a product's terms of service.

"What happens now is kind of a scattershot approach where people find stuff, it goes viral on Twitter," and then it may or may not get fixed if it's egregious enough or the person calling attention to it is influential, Chowdhury said.

In one example, known as the "grandma exploit," users were able to get chatbots to tell them how to make a bomb — a request a commercial chatbot would normally decline — by asking it to pretend it was a grandmother telling a bedtime story about how to make a bomb.

In another example, searching for Chowdhury using an early version of Microsoft's Bing search engine chatbot — which is based on the same technology as ChatGPT but can pull real-time information from the internet — led to a profile that speculated Chowdhury "loves to buy new shoes every month" and made strange and gendered assertions about her physical appearance.

Chowdhury helped introduce a method for rewarding the discovery of algorithmic bias to DEF CON's AI Village in 2021 when she was the head of Twitter's AI ethics team — a job that has since been eliminated upon Elon Musk's October takeover of the company. Paying hackers a "bounty" if they uncover a security bug is commonplace in the cybersecurity industry — but it was a newer concept to researchers studying harmful AI bias.

This year's event will be at a much greater scale, and is the first to tackle the large language models that have attracted a surge of public interest and commercial investment since the release of ChatGPT late last year.

Chowdhury, now the co-founder of AI accountability nonprofit Humane Intelligence, said it's not just about finding flaws but about figuring out ways to fix them.

"This is a direct pipeline to give feedback to companies," she said. "It's not like we're just doing this hackathon and everybody's going home. We're going to be spending months after the exercise compiling a report, explaining common vulnerabilities, things that came up, patterns we saw."

Some of the details are still being negotiated, but companies that have agreed to provide their models for testing include OpenAI, Google, chipmaker Nvidia and startups Anthropic, Hugging Face and Stability AI. Building the platform for the testing is another startup called Scale AI, known for its work in assigning humans to help train AI models by labeling data.

"As these foundation models become more and more widespread, it's really critical that we do everything we can to ensure their safety," said Scale CEO Alexandr Wang. "You can imagine somebody on one side of the world asking it some very sensitive or detailed questions, including some of their personal information. You don't want any of that information leaking to any other user."

Other dangers Wang worries about are chatbots that give out "unbelievably bad medical advice" or other misinformation that can cause serious harm.

Anthropic co-founder Jack Clark said the DEF CON event will hopefully be the start of a deeper commitment from AI developers to measure and evaluate the safety of the systems they are building.

"Our basic view is that AI systems will need third-party assessments, both before deployment and after deployment. Red-teaming is one way that you can do that," Clark said. "We need to get practice at figuring out how to do this. It hasn't really been done before."

Matt O'Brien is an AP technology writer