Large language models (LLMs) such as ChatGPT and GPT-4 have shown impressive performance in complex reasoning tasks. However, it is difficult to know whether the models are reasoning based on deep understandings of truth and logic, or leveraging their memorized patterns in a relatively superficial way. In this work, we explore testing LLMs' reasoning by engaging with them in a debate-like conversation, where given a question, the LLM and the user need to discuss to make the correct decision starting from opposing arguments. Upon mitigating the Clever Hans effect, our task requires the LLM to not only achieve the correct answer on its own, but also be able to hold and defend its belief instead of blindly believing or getting misled by the user's (invalid) arguments and critiques, thus testing in greater depth whether the LLM grasps the essence of the reasoning required to solve the problem. Across a range of complex reasoning benchmarks spanning math, commonsense, logic and BIG-Bench tasks, we find that despite their impressive performance as reported in existing work on generating correct step-by-step solutions in the beginning, LLMs like ChatGPT cannot maintain their beliefs in truth for a significant portion of examples when challenged by oftentimes absurdly invalid arguments. Our work points to danger zones of model alignment, and also suggests more careful treatments and interpretations of the recent findings that LLMs can improve their responses based on feedback.

Don't miss The Business of Marketing podcast featuring leading c-suite executives sharing insights on the importance of leveraging the intersectionality of marketing, finance, technology, HR and the boardroom to drive business growth.

The fall season brings many familiar favorites. It's common nowadays to see notifications from healthcare organizations on the local news alongside email reminders from employers about annual flu shots. If anything, it's a normal occurrence--perhaps anticipated, alongside ads for new pumpkin spice- flavored consumables. But even with careful preparation, healthcare professionals often work behind the curve to track the progress of reported flu outbreaks. Numerous factors are at play.

Our mission at The Weather Company is to help people make informed decisions in the face of weather, and there is no better time to connect with consumers than during flu season. Today, The Weather Company introduces Flu Insights with Watson in partnership with CVS, a new feature within The Weather Channel app for iOS and Android that leverages artificial intelligence (AI) and machine learning to arm consumers across the country with critical information to help them prepare for flu season. For the first time, Flu Insights with Watson includes a notification that is triggered during key moments like the onset of flu season, confirmed cases of flu outbreak or increased risk conditions. Users then have access to an industry-first 15-day flu forecast that displays the risk ranging from low to high, according to ZIP code. The experience also includes illness prevention tips and the latest flu reports from the CDC to provide users with actionable insights.

Smart speakers are becoming more and more common inside of homes, offering a convenient way to get the weather, manage your calendar, and answer any questions you may have. But this year, your Amazon Echo can also help you prepare for the upcoming flu season. Conceptualized and developed by Seattle Children's Hospital and Boston Children's Hospital, the Flu Doctor skill provides a convenient way to educate yourself (and your family) about the flu vaccine. "We know that search is increasingly going to be voice-enabled and we know increasingly more and more of us are incorporating smart speakers into our lives," Dr. Wendy Sue Swanson, general pediatrician and Chief of Digital Innovation at Seattle Children's Hospital, tells Reviewed. "The benefit of Flu Doctor is to learn more about the flu in your home, in a way that maybe you hadn't before using Alexa."

Ever since Facebook finally admitted to having a fake news problem, it's been trying to fix it. It hired thousands of people to help block fake ads, pledged to work with third-party fact-checking organizations and is busy building algorithms to detect fake news. But even as it attempts to fight back against fraudulent ads and made-up facts, another potential fake news threat looms on the horizon: Artificially generated fake video. Motherboard recently uncovered a disturbing new trend on Reddit, where users create AI-generated pornographic clips by swapping other people's faces onto porn stars. The outlet first reported on the phenomenon a month ago when Reddit user "deepfakes" posted a video of Gal Gadot's face swapped onto a porn star's body (he's since created more fake porn with other celebrities).

Modes of automation or artificial intelligence (AI) have been used by organizations for years. The concept goes back to the 1950s, according to Brian Kalis, managing director of digital health at Accenture, a business management consulting firm based in Boston. AI in healthcare goes beyond IBM Watson. In this e-guide, discover 4 uses for AI in healthcare – particularly how it can help improve patient engagement – and whether we can overcome security and interoperability concerns surrounding the technology. You forgot to provide an Email Address.