The Future of AI
LLMs are gaining traction as the workforce across domains is exploring artificial intelligence and automation to plan their operations and make crucial decisions. Generative and Foundational models are thus relied on for multi-step reasoning tasks to achieve planning and execution at par with humans. Although this aspiration is yet to be achieved, we require extensive … Read more
When we began studying jailbreak evaluations, we found a fascinating paper claiming that you could jailbreak frontier LLMs simply by translating forbidden prompts into obscure languages. Excited by this result, we attempted to reproduce it and found something unexpected.
The German philosopher Fredrich Nietzsche once said that “invisible threads are the strongest ties.” One could think of “invisible threads” as tying together related objects, like the homes on a delivery driver’s route, or more nebulous entities, such as transactions in a financial network or users in a social network. Computer scientist Julian Shun studies … Read more
Technologies Published 18 September 2024 Authors Eli Collins New video generation technology in YouTube Shorts will help millions of people realize their creative vision Artificial intelligence (AI) technologies for generating creative content are improving rapidly, but seamless ways of using them still aren’t widely available. We’re changing that, and making these incredible technologies more easily … Read more
The path to AI isn’t a sprint – it’s a marathon, and businesses need to pace themselves accordingly. Those who run before they have learned to walk will falter, joining the graveyard of businesses who tried to move too quickly to reach some kind of AI finish line. The truth is, there is no finish … Read more
In the fast-paced world of digital marketing, creating SEO-optimized content can often feel like a balancing act between creativity and technical precision. But what if there was a tool that could ease this process, without sacrificing quality? Enter the world of AI SEO writers—an emerging technology that’s reshaping the way we approach content writing for … Read more
Welcome to our weekly roundup of human-crafted AI news. This week OpenAI handed out API goodies. California’s AI safety bill got killed. And NVIDIA surprised us with a powerful open model. Let’s dig in. Here come the agents OpenAI didn’t announce any new models (or Sora) at its Dev Day event, but developers were excited … Read more
Sample language model responses to different varieties of English and native speaker reactions.
ChatGPT does amazingly well at communicating with people in English. But whose English?
Only 15% of ChatGPT users are from the US, where Standard American English is the default. But the model is also commonly used in countries and communities where people speak other varieties of English. Over 1 billion people around the world speak varieties such as Indian English, Nigerian English, Irish English, and African-American English.
Speakers of these non-“standard” varieties often face discrimination in the real world. They’ve been told that the way they speak is unprofessional or incorrect, discredited as witnesses, and denied housing–despite extensive research indicating that all language varieties are equally complex and legitimate. Discriminating against the way someone speaks is often a proxy for discriminating against their race, ethnicity, or nationality. What if ChatGPT exacerbates this discrimination?
To answer this question, our recent paper examines how ChatGPT’s behavior changes in response to text in different varieties of English. We found that ChatGPT responses exhibit consistent and pervasive biases against non-“standard” varieties, including increased stereotyping and demeaning content, poorer comprehension, and condescending responses.
A recent award from the U.S. Defense Advanced Research Projects Agency (DARPA) brings together researchers from Massachusetts Institute of Technology (MIT), Carnegie Mellon University (CMU), and Lehigh University (Lehigh) under the Multiobjective Engineering and Testing of Alloy Structures (METALS) program. The team will research novel design tools for the simultaneous optimization of shape and compositional … Read more
Research Published 26 September 2024 Authors Anna Goldie and Azalia Mirhoseini Our AI method has accelerated and optimized chip design, and its superhuman chip layouts are used in hardware around the world In 2020, we released a preprint introducing our novel reinforcement learning method for designing chip layouts, which we later published in Nature and … Read more