On July 19, Penn Medicine experienced a major disruption when a global IT outage from CrowdStrike affected 50,000 of their computers, leaving staff unable to use electronic medical records and greeted with the "blue screen of death." Executives, awakened by early morning calls, held an emergency Zoom meeting to manage the crisis, which led to the cancellation of most nonessential procedures and a switch to paper records for documentation. Critical imaging machines remained operational on a different system. Penn Medicine is extending clinic hours and considering a shift to iOS devices to mitigate future disruptions.

Inside Penn Medicine during the global IT outage Becker's Hospital Review

The McKinsey article explores the evolving role of generative AI (gen AI) from providing information to executing complex workflows as "agents," akin to virtual coworkers. These gen AI-powered agents leverage foundation models to adapt to new scenarios, use natural language for task instructions, and integrate seamlessly with digital ecosystems, potentially automating intricate workflows that previously required human intervention. Early investments by tech giants like Google, Microsoft, and OpenAI signal the rise of agentic systems, which could streamline processes such as loan underwriting, code modernization, and online marketing. However, the technology remains nascent, warranting careful strategic planning and human oversight to mitigate risks and maximize its potential in business applications.

Why agents are the next frontier of generative AI McKinsey Quarterly

Meta’s Llama 3.1 family of large language models (LLMs) presents significant advantages for enterprises by offering open weights, improved accuracy, and cost-saving opportunities. Analysts note that these models empower businesses to develop custom AI solutions, avoid vendor lock-in, and leverage high-performance capabilities without expensive proprietary licenses. However, the high deployment costs of the 405 billion-parameter variant might be prohibitive for small to medium enterprises, although cloud-based API solutions could mitigate this. The open nature of Llama 3.1 poses a competitive challenge to proprietary LLM providers, potentially driving down costs and fostering innovation across the AI industry.

Why Meta’s Llama 3.1 is a boon for enterprises and a bane for other LLM vendors InfoWorld

1. **EU, UK, and US Unite in GenAI Concerns**: Competition authorities from the US, EU, and UK have collectively voiced their worries about potential antitrust issues within the realm of generative AI, such as market concentration and anti-competitive practices. Top officials emphasized the necessity for swift measures to manage the fast-paced developments in AI technology. Highlighted risks include control over critical resources, entrenchment of market power, and potentially harmful partnerships. The collaborative stance indicates a unified approach to oversight, with likely increased scrutiny on AI-related mergers and business practices in the upcoming months.

2. **EU’s AI Regulation Sparks Concern From Meta**: Meta, Facebook's parent company, has expressed concern over the European Union's stringent AI regulations, fearing they might isolate Europe from advanced AI services. Meta's deputy privacy officer, Rob Sherman, emphasized that regulatory clarity is crucial to prevent hindering the deployment of cutting-edge technologies in the EU. This sentiment reflects the broader challenge of balancing innovation with responsible AI governance. Meta's compliance with recent EU requests has already delayed AI assistant rollouts, highlighting potential impacts on Europe's competitive edge in the global AI market.

3. **UK Labour Government Treads Carefully on AI Regulation**: Under Prime Minister

EU, UK and US Unite in GenAI Concerns PYMNTS