A new study conducted by researchers at MIT and elsewhere has found large language models (LLMs) can be used to predict the future as well as humans can, reports Chris Stokel-Walker for Fast Company. “Accurate forecasting of future events is very important to many aspects of human economic activity, especially within white collar occupations, such as those of law, business and policy,” says postdoctoral fellow Peter S. Park.
Prof. Daniela Rus, director of CSAIL, speaks with The Economist’s Babbage podcast about the history and future of artificial neural networks and their role in large language models. “The early artificial neuron was a very simple mathematical model,” says Rus. “The computation was discrete and very simple, essentially a step function. You’re either above or below a value.”
Wardah Inam SM '12, PhD '16 founded Overjet, an AI platform that helps dentists “diagnose diseases from scans and other data,” reports Saritha Rai for Bloomberg. “Dentistry was more art than science, and I wanted to bring technology and AI to help dentists make objective decisions,” says Inam. “We began building and then improving our AI systems with tens of millions of pieces of data, including X-rays, historical information, dentist notes, and periodontal charts.”
Senior Lecturer Luis Videgaray speaks with Government Technology reporter Nikki Davidson about concerns facing emerging AI programs and initiatives. Videgaray underscores the importance of finding vendors, "who are willing to protect the data in a way that is appropriate and also provides the state or local government agency with the required degree of transparency about the workings of the model, the data that was used for training and how that data will interact with the data supplied by the customer.”
Prof. Philip Isola and Prof. Daniela Rus, director of CSAIL, speak with Associated Press reporter Matt O’Brien about AI generated images and videos. Rus says the computing resources required for AI video generation are “significantly higher than for still image generation” because “it involves processing and generating multiple frames for each second of video.”
Boston Globe reporter Michael Silverman spotlights the 18th MIT Sloan Sports Analytics Conference. The conference focused on a, “diverse array of heady topics such as artificial intelligence, the globalization of soccer, the next phase of sports ownership, the evolutional of poker strategy,” writes Silverman, noting that “nearly every conversation on stage seemed to circle back to a shared belief that the momentum already carrying women’s sports is on the verge of a new surge.”
Prof. Daniela Rus, director of CSAIL, speaks with Boston Globe reporter Evan Sellinger about her new book, “The Heart and the Chip: Our Bright Future With Robots,” in which she makes the case that in the future robots and humans will be able to team up to create a better world. “I want to highlight that machines don’t have to compete with humans, because we each have different strengths. Humans have wisdom. Machines have speed, can process large numbers, and can do many dull, dirty, and dangerous tasks,” Rus explains. “I see robots as helpers for our jobs. They’ll take on the routine, repetitive tasks, ensuring human workers focus on more complex and meaningful work.”
Prof. Stuart Madnick speaks with CNBC reporter Kevin Williams about how the rise of generative AI technologies could lead to cyberattacks on physical infrastructure. “If you cause a power plant to stop from a typical cyberattack, it will be back up and online pretty quickly,” Madnick explains, “but if hackers cause it to explode or burn down, you are not back online a day or two later; it will be weeks and months because a lot of the parts in these specialized systems are custom made.”
Senior Lecturer Guadalupe Hayes-Mota writes for Forbes about the ways AI is reshaping drug development. “In the next three years, we can anticipate a more streamlined, efficient and cost-effective drug development process, ultimately leading to faster access to life-saving drugs for patients worldwide,” Hayes-Mota writes. “This is not just an evolution; it is a revolution in healthcare powered by the intelligence of machines.”
Researchers at MIT and elsewhere have developed a machine-learning model that can identify which drugs should not be taken together, reports Politico. “The researchers built a model to measure how intestinal tissue absorbed certain commonly used drugs,” they write. “They then trained a machine-learning algorithm based on their new data and existing drug databases, teaching the new algorithm to predict which drugs would interact with which transporter proteins.”
Quartz reporter Michelle Cheng spotlights a working paper by Prof. David Autor which shows that “AI could enable more workers to perform higher-stakes, decision-making tasks that are currently relegated to highly-educated workers such as doctors and lawyers.” As Autor explains, “in essence, AI used well can assist with restoring the middle-skill, middle-class heart of the US labor market that has been hollowed out by automation and globalization.”
MIT researchers have developed a new technique “that could allow most large language models (LLMs) like ChatGPT to retain memory and boost performance,” reports Tony Ho Tran for the Daily Beast. “The process is called StreamingLLM and it allows for chatbots to perform optimally even after a conversation goes on for more than 4 million words,” explains Tran.
In an article co-authored for The Economist, Senior Lecturer Donald Sull explores the impact of artificial intelligence and large language models (LLMs) on corporate company culture. “Leaders who do adopt AI for cultural insights can use these to make their employees happier, lower the odds of reputational disasters and, ultimately, boost their profits,” writes Sull. “Measurement is not the only piece of the ‘successful culture’ puzzle, but it is a crucial one. Culture has always been an enigma at the heart of organizational performance: undoubtedly important, but inscrutable. With AI, meaningful progress can be made in deciphering it.”
Prof. David Rand speaks with VOA News about the potential impact of adding watermarks to AI generated materials. “My concern is if you label as AI-generated, everything that’s AI-generated regardless of whether it’s misleading or not, people essentially are going to stop really paying attention to it,” says Rand.
Researchers at MIT and elsewhere have found that while AI systems could help doctors come to the right diagnosis more often, the diagnostic gains aren’t always distributed evenly, with more improvements tied to patients with lighter skin, report Daniel Payne, Erin Schumaker, and Ruth Reader for Politico. “AI could be a powerful tool to improve care and potentially offer providers a check on their blindspots," they write. "But that doesn’t mean AI will reduce bias. In fact, the study suggests, AI could cause greater disparities in care.”