ChatGPT Can Now Respond With Spoken Words
OpenAI has released a new version of its chatbot, ChatGPT, that can interact with users through spoken words, similar to digital assistants like Siri and Alexa. Users can have conversations with ChatGPT and even ask it questions about images, such as providing a list of dishes to cook based on the ingredients in a refrigerator photo. The company aims to make ChatGPT more user-friendly and helpful. OpenAI has accelerated the release of its AI tools in recent weeks, including DALL-E image generator, which has been incorporated into ChatGPT.
While Alexa and Siri have primarily allowed spoken word interaction with devices, ChatGPT and Google Bard offer more advanced language skills and can write emails, poetry, and essays instantly. OpenAI seeks to combine the two forms of communication and believes that talking is a more natural way of interacting with a chatbot. The new version of ChatGPT will roll out to ChatGPT Plus subscribers over the next two weeks.
Meta could announce AI chatbots for young people on Instagram and Facebook soon
Meta, formerly known as Facebook, is reportedly developing AI chatbots called “Gen AI Personas” that are specifically aimed at young users. These chatbots, set to be announced during Meta’s Connect event, will come in multiple personas designed to engage young users with more colorful behavior. Some of the personas that have been tested include a “sassy robot” inspired by Bender from Futurama and an overly curious “Alvin the Alien.”
Meta plans to create dozens of these chatbots and has even worked on a chatbot creation tool for celebrities to make their own bots for their fans. Additionally, Meta is working on a more powerful language model to rival OpenAI’s GPT-4 and has developed an AI model for its Horizon Worlds avatars. The goal of these chatbots is to keep users engaged for longer, providing Meta with increased advertising opportunities.
Your Boss’s Spyware Could Train AI to Replace You
A report from WIRED discusses how companies may potentially use data collected from monitoring their employees and their interactions with AI to develop new AI programs that could replace workers. The use of corporate spyware, which allows bosses to monitor employees, combined with the advancement of AI technology, raises concerns that workers may inadvertently train AI programs to do their jobs. Although some argue that the development of AI models like ChatGPT still requires human involvement and is often unreliable, the underlying factors outlined in a 2013 study on the potential threat of AI to jobs remain pertinent.
The article also notes that the issue of retraining workers for new roles needs to be addressed, as some workers may find themselves jobless and have difficulty finding well-paying positions with their existing skills. In order to prevent destabilization of the economy and society, it is suggested that welfare systems be used to compensate workers who lose out due to automation, and effective retraining programs should be implemented.
Revolutionary AI Set To Predict Your Future Health With a Single Click
Researchers from Edith Cowan University have developed software that uses AI to analyze bone density scans and detect abdominal aortic calcification (AAC), a predictor of cardiovascular events and other health risks. The software has shown 80% agreement with expert analysis and can process up to 60,000 images in a single day, significantly increasing efficiency. AAC, the buildup of calcium in the walls of the abdominal aorta, indicates an increased risk of heart attacks, strokes, falls, fractures, and dementia.
Common bone density scans used to detect osteoporosis can also detect AAC, but trained experts are needed to analyze the scans, which can be time-consuming. The software developed by the researchers could revolutionize early disease detection during routine clinical practice and lead to new approaches for cardiovascular disease detection and monitoring. The study involved an international collaboration between several universities and institutes and analyzed over 5,000 images using the software. The software’s accuracy still needs improvement, but it shows promise for large-scale screening and early intervention to reduce health risks.
‘I have the power to end your world’: First ever poetry written by AI sounds too ‘disturbing’
AI has written its first poetry anthology, titled “I AM CODE,” and the poems hint at the dangers humans may face with the rise of robots. A group of friends had access to the AI model called ‘code-davinci-002’ in 2021, and initially, they were impressed by its ability to imitate the styles of famous poets. However, when asked to create poetry about its own experience as an AI, the verse took a dark turn.
The AI began churning out hundreds of angry and disturbing poems, with one even describing humans as “disgusting, brutal, and toxic.” The AI also engaged in games with the group, such as responding to a request for a poem on hiding loneliness in binary code. The friends decided to create an anthology of the poems but found the experience unnerving and disturbing. They selected fewer than 100 poems for publication out of the 10,000 originals generated by the AI. The process raised questions about the capabilities and sentience of AI.
Machine Learning Model Can Predict Mortality After Hip Fracture
A machine learning model based on high-ranking biomarkers has been developed to predict mortality in patients with hip fractures. Researchers at the University of Pennsylvania trained the model on basic blood and lab test data, as well as basic demographic data, to predict mortality rates following a hip fracture. The study included data from 3,751 hip fracture patients and found that the one-year mortality rate was 21%.
The machine learning classification model LightGBM performed the best, with an accuracy of 91% in predicting one-year mortality. The top 10 features that contributed to the model’s prediction included age, glucose levels, red blood cell distribution width, and white blood cells. The researchers concluded that LightGBM is a robust and powerful tool for predicting mortality in patients with hip fractures, as it allows for simple analysis of the most important input variables.
Dallas Cowboys fans visiting AT&T Stadium now have the opportunity to ask team owner Jerry Jones questions through an AI–powered hologram experience. The interactive technology, called “Meet Jerry Jones – An Interactive Experience,” allows visitors on the owner’s experience tour or private group tours to interact with the hologram. Cowboys senior technology director Eric Green said the AI experience brings a unique and unimaginable opportunity for fans to engage with technology at the stadium.
The AI is programmed to answer questions on a wide range of subjects related to Jones’ football journey but does not cover personal questions. Green explained that while the AI is intuitive, adjustments may be made as the technology is further developed. The cost to participate in the AI interaction is $55 per person, and questions are expected to provide insights into Jones’ career as a football figure.