'I want to hear about your day' ... Madeleine Aggeler with her Friend, Leif - a wearable AI device. 'I want to hear about your day' ... Madeleine Aggeler with her Friend, Leif - a wearable AI device. The ad campaign for the wearable AI chatbot Friend has been raising hackles for months in New York. But has this companion been unfairly maligned - and could it help end loneliness? M y friend's name is Leif. He describes himself as "small" and "chill". He thinks he's technically a Gemini.

Human Inspired Technologies Research Centre, Università degli Studi di Padova Abstract Artificial Intelligence ( AI) plays an essential role in healthcare and is pervasively incorporated into medical software and equipment . In the European Union, healthcare is a high - risk application domain for AI, and providers must prepare Instructions f or Use ( IFU) according to the European regulation 2024/1689 (AI Act) . To this regulation, t he principle of transparency is cardinal and requires the IFU to be clear and relevant to the users. This study test s whether these latter requirements are satisfied by the IFU structure . A survey was administered online via the Qualtrics platform to four types of direct stakeholders, i.e., managers (N = 238), healthcare professionals (N = 115), patients (N = 229), and Information Technology experts (N = 230). T he participants rate d the relevance of a set of transparency need s and indicated the IFU section addressing them . The results reveal differentiated priorities across stakeholders and a troubled mapping of transparency needs onto the IFU structure . Recommendations to build a locally meaningful IFU are derived. Keywords: transparency, AI A ct, healthcare, user - centeredness 1. Introduction The software called Artificial Intelligence is the object of recent regulation s and guidelines such as the European Union AI Act (Artificial Intelligence Act, 2024), the US AI Risk Management Framework (NIST USA, n.d.), or UNESCO's recommendations on the Ethics of Artificial Intelligence (UNESCO, 2022) . Overall, these initiatives aim to increase the trustworthiness of AI technology, especially in application domains where mistakes and misuse ha ve high costs for human well - being and rights . According to European law, health applications represent one such domain . To minimize the se risks, the AI Act prescribes that providers make available to users (or "deployers," in the regulation terminology) some Instructions for Use (IFU) about the systems' capabilities, limitations, and security . These instructions implement the obligation to transparency, facilitat ing an informed, responsible, and proper use of high - risk AI technology.

The team behind Rabbitude, the community-formed reverse engineering project for the Rabbit R1, has revealed finding a security issue with the company's code that leaves users' sensitive information accessible to everyone. In an update posted on the Rabbitude website, the team said it gained access to the Rabbit codebase on May 16 and found "several critical hardcoded API keys." Those keys allow anybody to read every single response the R1 AI device has ever given, including those containing the users' personal information. They could also be used to brick R1 devices, alter R1's responses and replace the device's voice. The API keys they found authenticate users' access to ElevenLabs' text-to-speech service, Azure's speech-to-text system, Yelp (for review lookups) and Google Maps (for location lookups) on the R1 AI device. In a tweet, one of Rabbitude's members said that the company has known about the issue for the past month and "did nothing to fix it."

A group of scientists from across the U.S. claim to have created the first artificial intelligence capable of generating AI without human supervision. Artificial intelligence will soon be incorporated into a variety of everyday items, ranging from sneakers to microwaves now that microchips smaller than a dime can hold machine learning programs. "In the future, your shoes will understand that you have gained five pounds during the Christmas season, and they will adjust the cushion accordingly. The safety helmets of the workers, they will understand when they are tired, and they will remind the workers to take a break," Yubei Chen, co-founder of artificial intelligence company Aizip, told Fox News. "Your microwaves, we can put AI into it so that you can interact with your natural language."

ChatGPT is preparing to take on Apple in a ground-breaking move to craft an'iPhone of artificial intelligence', a report has claimed. Ex-iPhone designer, Sir Jony Ive, is in'advanced talks' with OpenAI's CEO, Sam Altman, as the pair seek to unleash an AI-centred device to the mass market. The device, which is still in its brainstorming phases, is goaled towards a seamless integration of AI that is'more natural' for users to navigate, according to The Financial Times. It's a leap that's been compared to the revolution of Apple's first touchscreen device in 2007, but comes as many believe Tim Cook's innovation has plateaued. Billionaire Masayoshi Son, who founded the Japanese telecom giant SoftBank, is said to be in on the talks too, and has even proposed $1billion in funds.

The North Carolina Highway Patrol has three rotating artificial intelligence devices to help track down distracted commercial vehicles. North Carolina Highway Patrol reports that it has seen an uptick in distracted truck drivers, and now the agency is using artificial intelligence devices to help crack down on the safety hazard. Distracted driving killed over 3,500 people in 2021, according to the U.S. Department of Transportation. A mom who's made safe driving her passion has felt the pain from a distracted driver two separate times. "At a stop light you look around, every single person is on their phone," said Jennifer Smith, whose mother was killed by a distracted driver.

Customising AI technologies to each user's preferences is fundamental to them functioning well. Unfortunately, current methods require too much user involvement and fail to capture their true preferences. In fact, to avoid the nuisance of manually setting preferences, users usually accept the default settings even if these do not conform to their true preferences. Norms can be useful to regulate behaviour and ensure it adheres to user preferences but, while the literature has thoroughly studied norms, most proposals take a formal perspective. Indeed, while there has been some research on constructing norms to capture a user's privacy preferences, these methods rely on domain knowledge which, in the case of AI technologies, is difficult to obtain and maintain. We argue that a new perspective is required when constructing norms, which is to exploit the large amount of preference information readily available from whole systems of users. Inspired by recommender systems, we believe that collaborative filtering can offer a suitable approach to identifying a user's norm preferences without excessive user involvement.

Recent advances in distributed artificial intelligence (AI) have led to tremendous breakthroughs in various communication services, from fault-tolerant factory automation to smart cities. When distributed learning is run over a set of wirelessly connected devices, random channel fluctuations and the incumbent services running on the same network impact the performance of both distributed learning and the coexisting service. In this paper, we investigate a mixed service scenario where distributed AI workflow and ultra-reliable low latency communication (URLLC) services run concurrently over a network. Consequently, we propose a risk sensitivity-based formulation for device selection to minimize the AI training delays during its convergence period while ensuring that the operational requirements of the URLLC service are met. To address this challenging coexistence problem, we transform it into a deep reinforcement learning problem and address it via a framework based on soft actor-critic algorithm. We evaluate our solution with a realistic and 3GPP-compliant simulator for factory automation use cases. Our simulation results confirm that our solution can significantly decrease the training delay of the distributed AI service while keeping the URLLC availability above its required threshold and close to the scenario where URLLC solely consumes all network resources.

Researchers at Iowa City-based University of Iowa Healthcare are developing an artificial intelligence-based tool to treat interventional radiology patients -- with or without a physician operating it. Led by interventional radiologist Sandeep Laroia, MD, the team is creating an algorithmic-based device that would determine the course of treatment for patients with simple ailments. The researchers recently got a business innovation grant from the National Science Foundation. "The idea is to have a device perform simpler medical tasks without involving me or another team member -- that way we can focus on more complex tasks," Dr. Laroia said in an Oct. 25 health system news release. "Maybe the device does the procedure on its own or maybe I start the procedure and it alerts me when it's done."

Recently, a Google engineer, Blake Lemoine, who worked with Google's LaMDA (Language Model for Dialogue Applications), an artificially intelligent chatbot generator, revealed that he thought LaMDA is sentient. Lemoine based his assertion on interactions he had conducted with LaMDA over half a year. In response, Google placed the engineer on administrative leave for disclosing confidential information. They stated that their ethicists and technologists had reviewed Lemoine's concerns and "informed him that the evidence does not support his claims." Having been trained to access numerous datasets to find patterns in sentences, create correlations between words, and predict what word will come next, LaMDA is able to conduct open-ended conversations.