Stay ahead of the headlines

Jun 21, 2024·theregister.com

A global survey by Nash Squared reveals that the anticipated transformative impact of Generative AI (GenAI) on the workplace has not yet materialized, despite significant vendor valuations and lofty societal predictions. Interviews with 322 tech leaders globally indicate that while 51% of organizations have implemented GenAI to varying degrees, and 21% to a greater extent, the practical uses have mostly been limited to personal productivity tools like document creation and research. Many organizations are still in the experimental phase, with 39% reporting little to no impact from the technology. The survey also highlights that concerns about misuse, budgetary constraints, and an unproven business case for mass investment are hindering broader adoption. Additionally, the effect on jobs has been minimal, with only 1% of roles being replaced by GenAI and 10% seeing changes in job roles.

Jun 21, 2024·The Hustle

A recent report by Harvard Business School, Imperial College Business School, and the German Institute of Economic Research highlights a decline in demand for online freelance jobs since the debut of ChatGPT in November 2022. The findings indicate a 30% drop in demand for writers, 20% for coders, 17% for graphic designers, and 13% for social media and data entry roles. Despite this decline, pay rates have minimally increased. The rise of AI has also spurred disputes on freelance platforms over the use of AI-generated content, with some freelancers now offering services to edit AI-created work. The landscape suggests AI is taking over simpler tasks traditionally handled by freelancers, prompting concerns over fair compensation and job quality.

Jun 20, 2024·Fortune

Amazon's attempt to transform Alexa with generative AI has stumbled, according to former employees. Despite a promising demo in September 2023, the new Alexa has yet to launch widely. Structural dysfunctions, technological constraints, and prioritization of other projects, such as AWS's Bedrock, have hindered progress. Insufficient data and outdated chips further impede development. As rival companies like OpenAI and Apple capture the generative AI market, Amazon's vision of making Alexa a leading conversational agent is in jeopardy, making it difficult for the digital assistant to catch up to its competitors.

Jun 19, 2024·Forbes

The article discusses the significant challenges faced by the 53 million unpaid family caregivers in the U.S., highlighting the health and financial stresses that contribute to high levels of burnout and deteriorated well-being. It introduces TCARE, Inc., a caregiver support platform co-founded by Ali Ahmadi, designed to mitigate these challenges using a combination of AI and personalized support from trained care specialists. The TCARE system leverages AI to assess caregiver needs, customize care plans, and connect caregivers to relevant resources, showing positive outcomes such as reduced depression and delayed need for institutional care. The piece emphasizes the importance of providing adequate support to caregivers for both their benefit and the care recipients.

Jun 19, 2024·dermatologyadvisor

The American College of Physicians (ACP) has issued a policy paper with ten recommendations on using artificial intelligence (AI) in healthcare. Published in the *Annals of Internal Medicine*, the paper emphasizes that AI tools should augment rather than replace clinical decision-making and adhere to medical ethics. The ACP calls for transparency, data privacy, and continuous improvement based on real-world feedback. The paper also stresses the need for training at all levels of medical education and advocates for a coordinated federal AI strategy to address biases and disparities. Additionally, the environmental impact of AI should be considered, and AI developers must be held accountable for their technology's performance.

Jun 19, 2024·hitconsultant.net

A study from Mount Sinai Hospital indicates that artificial intelligence (AI) can significantly enhance patient outcomes by predicting and preventing health decline. Published in *Critical Care Medicine*, the research involved over 2,700 adult patients split into two groups, with one group receiving real-time AI alerts about potential health deterioration. The intervention group saw a 43% higher likelihood of receiving escalated care and markedly reduced mortality rates. The study suggests that AI algorithms outperform traditional methods like the MEWS by enabling earlier interventions and potentially saving lives. Despite the study's early termination due to the COVID-19 pandemic, a simplified AI system is being used to monitor stepdown unit patients at Mount Sinai, continually improving through reinforcement learning.

Jun 19, 2024·techradar

Oracle, Microsoft, and OpenAI are collaborating to integrate the Microsoft Azure AI platform with Oracle Cloud Infrastructure (OCI) in response to increased demand for AI solutions. This alliance, partly driven by the integration of Apple Intelligence in new Apple OS updates, aims to extend AI capabilities and handle the expected rise in users. Apple's advancements in on-device processing and Private Cloud Compute are set to increase traffic for ChatGPT, a service used by over 100 million users monthly. Oracle's advanced AI infrastructure, including scalable Nvidia GPU capabilities, will support the development and training of large language models and other AI applications, making it a prime choice for leaders like OpenAI.

Jun 17, 2024·VentureBeat

In a recent development, Google’s Personal Health Large Language Model (PH-LLM), a specialized version of its Gemini LLM, has demonstrated superior performance in providing sleep and fitness advice compared to human experts. The PH-LLM, fine-tuned to analyze and interpret time-series data from wearable devices, was tested on its ability to answer health-related questions and make accurate predictions. Researchers found that the model consistently outperformed experienced professionals in the health and fitness domain, showcasing its potential to enhance personal health management through advanced AI technology.

Jun 17, 2024·Medical Economics

### Study: AI-powered Alerts Improve Patient Outcomes (June 14, 2024) A study conducted at Mount Sinai Hospital, published in Critical Care Medicine, revealed that AI-generated alerts significantly enhance patient care and outcomes. Patients whose care teams received real-time AI alerts about health deterioration were 43% more likely to get timely escalated care and had a lower mortality rate within 30 days. The research led by Dr. Matthew Levin demonstrates that AI-driven alerts outperform traditional methods in predicting clinical decline, facilitating earlier interventions and saving lives. Although the study faced interruptions due to the COVID-19 pandemic, the algorithm has been widely implemented across Mount Sinai's stepdown units, aiding clinical decision-making and improving patient safety through continuous data integration and reinforcement learning.

Jun 17, 2024·Forbes

The article "7 Paradoxes of Artificial Intelligence" by Joe McKendrick explores seven conflicting aspects of AI technology that complicate its implementation for business and IT leaders. These paradoxes include: AI reducing labor needs while increasing skill requirements, being complex yet simplifying application development, incurring high costs but aiding cost management, automating work while demanding creativity, benefiting advanced companies more than struggling ones, requiring large datasets yet alleviating data management, and possessing significant intelligence but failing at simple tasks. These contradictions highlight the nuanced and perplexing nature of AI adoption in the business world.

Jun 17, 2024·beckershospitalreview.com

Nurses are raising concerns about the integration of artificial intelligence (AI) into electronic health records (EHRs), arguing that the technology is ineffective and detracts from patient care. They report that AI-driven programs like automated nurse handoffs and patient classification systems frequently fall short of practical needs, citing examples where algorithms misjudge patient acuity or provide inaccurate warnings, like sepsis alerts. While companies like Epic and Oracle Health promote AI as a tool to enhance care and prevent adverse events, nurses insist that these systems fail to account for essential aspects of their workflow, such as patient education and compassionate care. Health systems like Kaiser Permanente and Keck Medicine are attempting to address these concerns by emphasizing that ultimate decisions are made by licensed nurses and that AI should assist rather than replace human judgment.

Jun 17, 2024·The Verge

Apple introduced Apple Intelligence at its WWDC event, a suite of AI tools for iPhone, iPad, and Mac, emphasizing both utility and data privacy. Apple Intelligence is designed to perform common AI tasks on-device, such as email drafting and call transcriptions, to ensure speed and privacy. For more complex tasks, the system reaches out to cloud servers in a secure manner. Notably, Apple uses its own AI models trained on publicly available data and licensed materials, rather than user data. The models are optimized for power efficiency and size to function effectively on Apple's devices. This dual approach aims to provide robust AI features while minimizing privacy concerns by keeping most data processing local and leveraging secure cloud computing for more intensive tasks. The success of this balance will be clearer once Apple Intelligence is fully deployed.

Jun 14, 2024·JMIR AI

The article "Forecasting Artificial Intelligence Trends in Health Care: Systematic International Patent Analysis" by Stan Benjamens, Pranavsingh Dhunnoo, Márton Görög, and Bertalan Mesko provides a comprehensive overview of AI and ML trends in health care through an international patent analysis. The study examined patents from major international patent offices between 2012 and 2022, highlighting a significant annual increase in AI health care patents, primarily from China. Key players identified included Ping An Medical and Healthcare Management Co Ltd and Siemens Healthineers, among others. The article emphasizes the importance of understanding these trends to guide regulators in developing appropriate guidelines for emerging technologies. An open-access database was also created for better accessibility to this information.

Jun 13, 2024·Apple

At the 2024 Worldwide Developers Conference, Apple introduced Apple Intelligence, a personal intelligence system integrated into iOS 18, iPadOS 18, and macOS Sequoia. Apple Intelligence features generative models designed to enhance everyday tasks, like refining text, summarizing notifications, and creating playful images. Key components include a ~3 billion parameter on-device language model and a larger server-based model leveraging Private Cloud Compute. The development adheres to Responsible AI principles, emphasizing user empowerment, representation, careful design, and privacy protection. Training involves a combination of licensed and public data, with rigorous data quality and safety protocols. Apple’s models outperform many competitors in user preference for helpfulness and safety, facilitated by advanced fine-tuning and optimization techniques.

Jun 12, 2024·cio.com

John Napoli, head of transformation at Guardian Life, is gearing up the company's AI initiatives by scaling up AI talent. Facing a tough labor market, Napoli is training existing employees and hiring strategically to meet business demand. This includes conventional data skills and the latest AI tools. He emphasizes the importance of a robust AI strategy, balancing building and buying tactics to create a skilled workforce. Orla Daly of Skillsoft adopts a similar approach, focusing on upskilling staff and training workers to integrate AI into workflows. Despite these efforts, AI talent remains scarce, slowing down AI adoption across enterprises. High competition for AI skills means many companies are left to train their own staff to fill critical roles.

Jun 12, 2024·Wired

OpenAI has released a research paper on reverse engineering AI models in response to criticism about the potential risks of their technology. This comes days after former employees accused the company of being reckless. The paper explains a method for understanding how AI models store certain concepts, which can help identify misbehaviors in systems like ChatGPT. The research also underscores recent internal turmoil, including the disbanding of OpenAI's "superalignment" team and the departure of notable figures. This new technique, demonstrated on GPT-4, aims to make AI models more interpretable and controllable, although further refinement is needed. The research aligns with similar efforts in the field to make AI systems safer and more transparent.

Jun 12, 2024·Health Data Management

The article discusses how data democratization can address significant challenges in healthcare by improving the accessibility and usability of vast amounts of unstructured data generated by healthcare institutions. Data silos, lack of interoperability, and the fragmentation of electronic health record (EHR) systems hinder efficient data use, delaying critical medical interventions. The integration of artificial intelligence (AI) and technologies like natural language processing (NLP) can help extract valuable information from unstructured data, enabling better patient care and operational efficiencies. Despite these advantages, the article notes that financial, operational, and technical barriers must be overcome to fully realize the benefits of data democratization and AI in healthcare.

May 7, 2024·Constellation Research

Dion Hinchcliffe discusses the evolving role of Chief Information Officers (CIOs) in navigating the complex landscape of Artificial Intelligence (AI) integration within enterprises. In 2024, CIOs face increasing pressure from corporate boards to leverage AI strategically and rapidly, despite the technology’s fast-paced evolution and the absence of one-size-fits-all solutions. The article emphasizes the necessity for CIOs to develop a diverse portfolio of AI models, referred to as an "AI model garden," to meet varied business needs effectively and ensure alignment with overarching technology strategies. This approach requires a robust ModelOps infrastructure to manage the lifecycle of AI models comprehensively, ensuring they deliver business value while adhering to governance and regulatory standards. The piece underscores the strategic importance of AI readiness and the critical role of CIOs in fostering a culture of innovation and ensuring the ethical use of AI across the enterprise.

Apr 10, 2024·Fierce Healthcare

UMass Chan Medical School and MITRE Corporation are collaborating to establish an assurance lab in Worcester, Massachusetts, for testing healthcare artificial intelligence (AI) technologies. Funded by a state pledge of $555,000 and $137,000 in private sector investments, the lab will support the testing of AI-enabled solutions to ensure they are safe and reliable for healthcare applications. This initiative aims to develop a framework for evaluating these technologies, foster collaboration and workforce training in the region, and contribute to the creation of a national network of AI assurance labs. The lab's evaluation process will encourage the responsible use of AI in healthcare by assessing products' features and intended uses, with the goal of establishing emerging national standards for digital health technologies.

Apr 10, 2024·healthsystemcio.com

In the article "Like Herding Cats": Making a Case for AI Assurance Labs, John Halamka, President of the Mayo Clinic Platform, highlights the collaborative efforts of academic medical centers, technology companies, and federal agencies in establishing AI assurance labs. These labs aim to bring structure and standardization to the development and deployment of AI algorithms in healthcare, responding to concerns over their reliability and safety. Following President Joe Biden's executive order on AI algorithm guidelines, this initiative seeks to create a network of both private and public assurance labs for the technical and ethical evaluation of AI technologies. This effort addresses the challenges of a fragmented approach to AI development, aiming to ensure that algorithms are beneficial and non-biased across diverse patient groups and healthcare settings.