Elon Musk's artificial intelligence startup, xAI, is set to open source Grok, its chatbot that competes with ChatGPT, amid Musk's lawsuit against OpenAI for veering away from its open-source foundation. Grok, launched for X premium subscribers, features real-time information access and eschews political correctness. Musk's legal action challenges OpenAI's partnership with Microsoft, alleging a deviation from its original commitment to make AI technology widely available and accusing it of becoming a closed-source entity focused on profit. This move sparks broader discussions within the tech community about the value and future of open-source AI research, drawing mixed reactions from notable figures and commentators.

Apple, Tesla, and the Dying Dream of Self-Driving Cars

Mar 12, 2024New York Magazine Intelligencer

In the article "The Self-Driving Car Bubble Has Popped" by John Herrman for Intelligencer, the focus is on Apple's discontinuation of its nearly decade-long endeavor to develop a self-driving car, known internally as Project Titan. Despite the recruitment of industry veterans and significant financial investment, the project was hindered by numerous challenges, including technological limitations and changing market dynamics, leading to its abrupt end. The broader context reveals a cooling enthusiasm for autonomous vehicles across the industry, marked by overoptimistic forecasts, high-profile failures, and a reevaluation of goals toward more achievable driving-assistance technologies. This shift underscores the complex realities facing the pursuit of fully autonomous vehicles, as companies grapple with both the technical and practical obstacles ahead.

HIMSS24: Microsoft, 16 health systems to operationalize AI under new partner network

Mar 12, 2024Fierce Healthcare

On March 11, 2024, a coalition comprising Microsoft, 16 health systems, and various health technology organizations announced the formation of the Trustworthy & Responsible AI Network (TRAIN). This initiative aims to establish and promote best practices and standards for responsible AI application in healthcare. TRAIN seeks to enhance patient care and safety by fostering trust in AI technologies, which are increasingly used for patient screening and administrative tasks. The network will focus on sharing AI best practices, creating a secure online portal for registering clinical AI, developing a national AI outcomes registry, and providing tools for measuring AI effectiveness in diverse populations. This collaboration builds on principles from the Coalition for Health AI (CHAI) and includes notable health systems and tech providers committed to ensuring equitable access to TRAIN's resources for organizations of all sizes.

The Coalition For Health AI Is A ‘Goo-Goo,’ Not Yet A Revolutionary

Mar 12, 2024Forbes

The Coalition for Health AI (CHAI), having amassed over 1,300 members, was launched with the aim of becoming the trusted source for AI best practices in health, emphasizing inclusivity and good governance reminiscent of historic "goo-goo" good government groups. Although partnered with notable stakeholders including Google and Microsoft, CHAI's ambitious goals raise questions about its ability to confront contentious AI uses in health care, such as secretive data collection practices, and whether it will ultimately prioritize technical standards over empowering patients to make their own health decisions. While the organization promises to revolutionize patient care through AI, its true commitment to transformative change versus technical fixes remains to be seen, mirroring the dual legacy of good governance advocates of the past.

Salesforce announces new AI tools for doctors

Mar 9, 2024CNBC

Salesforce introduced AI tools for healthcare: Einstein Copilot and Assessment Generation to ease administrative tasks and combat physician burnout. These tools automate appointment setting, patient data summarization, and digital health assessments, integrating medical data from various sources. Aimed at reducing clerical burden, these solutions are part of Salesforce's broader strategy to streamline healthcare data management, with availability set for summer and year-end.

GE HealthCare, Vanderbilt AI models predict immunotherapy responses among cancer patients

Mar 9, 2024FierceBiotech

GE HealthCare, in collaboration with Vanderbilt University Medical Center (VUMC), has developed artificial intelligence (AI) models capable of predicting cancer patients' responses to immunotherapies with 70% to 80% accuracy by analyzing diverse sets of data from over 2,200 patients. These models utilize electronic medical records, including demographic, genomic, tumor, cellular, proteomic, and imaging data, to forecast both efficacy outcomes and the likelihood of severe side effects such as hepatitis, colitis, and pneumonitis. This advancement aims to move beyond the current trial-and-error methodologies and reliance on novel biomarkers, providing a more precise, data-driven approach to cancer treatment. GE HealthCare is exploring the commercialization of these AI technologies to enhance patient care and support biopharma drug development, contingent on regulatory approvals.

CrowdStrike CEO says cybercriminals are leveling up: 'It’s going to be the battle of AI in the future'

Mar 7, 2024CNBC

In a recent discussion with CNBC's Jim Cramer, CrowdStrike CEO George Kurtz emphasized the growing threat posed by cybercriminals who are leveraging advanced generative artificial intelligence to conduct sophisticated attacks, regardless of their own technical expertise. Kurtz highlighted how generative AI is making complex cyberattack techniques accessible to less skilled adversaries, suggesting a future where the fight against cybercrime will increasingly hinge on opposing AI capabilities.

Microsoft, OpenAI warn nation-state hackers are abusing LLMs | TechTarget

Mar 7, 2024SearchSecurity

Microsoft and OpenAI report nation-state hackers using AI and LLMs like ChatGPT for cyberattacks but don't see an immediate enterprise threat. They emphasize the need for prepared security measures due to this activity. Research identified five groups globally exploiting LLMs for tech vulnerability research, geopolitical intelligence, and enhancing attack methods. Despite the increased use, no significant LLM-based attacks were found yet. The report also discusses AI's potential to enhance social engineering attacks and offers strategies for mitigation, emphasizing AI tool security and responsible use.

Nation states buying hacking tools from underground Russian cyber forums

Mar 7, 2024Computer Weekly

Nation states buy malware on Russian forums for data-wiping attacks, pretending to be cyber criminals for deniability. Russian forums, like Exploit existing since 2000, offer hacking tools and services, including AI and deep fake technologies. Check Point found evidence of nation state interest, particularly for wiper software, suggesting higher budgets and strategic use of cybercrime tools for espionage and sabotage, blending legitimate security tools to avoid detection. Russian cybercriminals, profiting from these forums, enforce strict membership with vetting, showcasing a rich ecosystem of hacking services, including ransomware as a service, while avoiding attacks on Russian-speaking countries to maintain immunity and promote Russian interests.

Report Uncovers Massive Sale of Compromised ChatGPT Credentials

Mar 7, 2024Hackread

Group-IB report: Dark web sales of ChatGPT credentials expose AI, ransomware, macOS vulnerabilities; Asia-Pacific major target, zero-day exploit sales up 70%, AI misuse rises; 225,000 ChatGPT credentials leaked, 130,000 hosts compromised, North America most ransomware-affected, macOS attacks multiply; emphasizes cybersecurity investment.

Claude 3 Is The Most Human AI Yet

Mar 7, 2024Every

Jim Simons excelled in trading using algorithms; Composer allows easy creation and testing of such strategies for anyone. Claude 3, from Anthropic, praised for its human-like, creative AI, surpasses GPT-4 in nuance despite technical benchmark similarities. Its unique warmth and high token accuracy hint at groundbreaking AI advancements and synthetic data use, challenging conventional model evaluations and AI competition strategies.

I worked exclusively in Vision Pro for a week—here’s how it went

Mar 7, 2024Arstechnica

Using Apple's Vision Pro offers a futuristic, albeit not essential, work experience, struggling with physical meetings and steep price.

Mark Cuban’s Cost Plus Drugs will begin manufacturing its own medications this week

Mar 7, 2024Fortune

Mark Cuban's Cost Plus Drugs starts manufacturing its medications in Dallas, aiming for affordable pricing by eliminating middlemen.

Elon Musk sues OpenAI for abandoning its mission to benefit humanity

Mar 6, 2024The Verge

Musk sues OpenAI, claiming profit-chasing and Microsoft partnership violated its non-profit, open-source agreement. Lawsuit seeks reversion to mission of AGI benefitting humanity. Musk's prior AI safety concerns noted. OpenAI's GPT-4 model allegedly excels at reasoning. Microsoft-appointed board members lack AI expertise, claims lawsuit, as OpenAI-Microsoft partnership undergoes regulatory examination for competition impacts. OpenAI declines comment.

AI-Generated Clinical Summaries Require More Than Accuracy

Mar 6, 2024JAMA

AI's potential in medicine, especially with LLMs for clinical summaries, poses challenges like accuracy, FDA oversight, and influences on decision-making, highlighting the need for careful consideration and regulation.

Dell says on-prem is the future of AI

Mar 5, 2024CIO Dive

Dell focuses on on-prem AI & generative AI to boost PC business amid declining growth, sees enterprise demand for AI-optimized servers, aims for AI integration in PCs and edge networks, faces production challenges due to chip shortages.

Amazon training chatbot for healthcare: Report

Mar 4, 2024Beckers Hospital Review

Amazon training AI chatbot to handle patient messages, administrative queries for One Medical, amidst telehealth surge and job cuts; intended to reduce healthcare provider's administrative burdens, improve response time.

LockBit Making A Comeback After The Massive Takedown

Mar 4, 2024CyberSecurityNews

Ransomware group LockBit reemerges on dark web within week after takedown by law enforcement, despite arrests linked to group. U.S. Department offers rewards for group members' identification or arrest, group reinstates servers, plans FBI statement. LockBit exploits CitrixBleed vulnerability, shows adaptability, resilience to law enforcement efforts. Group responsible for 21% of 2023 ransomware attacks.

What Does Virtual First Mean In Healthcare? - The Medical Futurist

Mar 1, 2024Medical Futurist

'Virtual first' in healthcare prioritizes remote medical services before in-person visits, blending digital and physical resources. It increases accessibility, efficiency, and cost-saving, despite potential disparities in care access. Adoption remains largely US-based but potential for global rollout exists.

I went for a walk with Gary Marcus, AI’s loudest critic

Feb 28, 2024Technology Review

Gary Marcus, an AI critic, discusses his concerns over AI's current state, including exploitation by tech companies, misuse of copyrighted materials, and the absence of common sense in AI systems. He believes AI policy and action are urgently needed to avoid making the same mistakes as with social media companies. Marcus has written a book, Taming Silicon Valley, about AI regulation and public involvement for responsible AI.

Stay ahead of the headlines

Trending now

AMA Asking For Stronger Safeguards on AI Mental Health Chatbots

Andrea Daugherty Appointed Chief Information and Digital Transformation Officer at ARMC

Craig Richardville Appointed Chief Digital and Information Officer at UF Health

Former AMA Exec Margaret Lozovatsky Named CDIO At Premier Health

AI & Machine Learning

Elon Musk to open-source AI chatbot Grok this week