Editor’s note: This is AI Impact, Newsweek’s weekly newsletter where each week, we will explore how business leaders are unlocking real value through artificial intelligence.
Tap or click here to get this newsletter delivered to your inbox.
Core Intelligence
Agility Has a Structure
Rivian CEO RJ Scaringe has spent years quietly rethinking how autonomous vehicle technology should be built, and in a recent conversation with Eileen Falkenberg-Hull, Newsweek’s Senior Editor for Autos, he described how Rivian has been developing a self-improving, end-to-end AI system designed to scale across vehicle generations.
Developed largely in stealth, the platform spans hardware, perception and a large foundation model that can be updated continuously through software, an approach that reflects Rivian’s focus on long-term architectural control rather than a roadmap driven primarily by incremental feature releases.
What makes Rivian’s advance notable is not just the technology itself, but what it reveals about how AI is changing the economics of innovation. AI compresses development cycles and rewards architectures that can evolve continuously, bypassing the multi-year product resets that still define much of the auto industry.
Startups designed for this environment can source talent differently, redefine roles more fluidly and make sustained bets without carrying the operational weight that still anchors many industrial incumbents.
For stalwarts, those operational constraints often translate into slower decision-making and R&D models optimized for a world that no longer exists.
That helps explain why partnerships are becoming a defining feature of the AI transition. Where IBM showed what internal reinvention looks like when leadership commits to rebuilding from within (redesigning workflows, retraining employees and embedding AI across the business), the Volkswagen–Rivian relationship highlights a different response: reaching outward to access agility that can’t easily be recreated inside a legacy structure.
For Volkswagen, that turn outward follows years of effort to centralize software development through its CARIAD unit, an effort that underscored how difficult it is for legacy organizations to move at startup speed.
In the AI age, flexibility is not just a cultural virtue, it’s a structural advantage. And as Scaringe’s perspective makes clear, the companies best positioned to move fast are often the ones least burdened by how innovation used to work.
You can watch the interview and read more here: Volkswagen Won’t Get Rivian’s Autonomous Vehicle and AI Tech.
AI Impact Awards & Summit
The Newsweek AI Impact Awards seek to identify and recognize uniquely innovative AI solutions that solve critical business problems in different industry segments, or significantly advance capabilities. Recognition comes not from ideas, but from measurable IMPACT on business operations.
Register now – Early bird deadline ends 12/19.
Prompt Injection
“AGI (artificial general intelligence) is much further away than you think. Given Meta’s $90B investment, OpenAI’s near trillion-dollar commitments, you’d think it’s right around the corner, winner-take-all. But it’s not. Large language models are trained on a significant portion of the internet plus all the digitized books and materials they can get their hands on. Each new model can cost $100M+ in compute and expensive Nvidia chips. But it’s all trained at once — a brain that has learned everything but is frozen in time. Time is one main reason why AGI is not upon us. Models don’t continuously learn. They don’t adapt to new information. They don’t understand that time and causality are inextricably interlinked. Without continuous learning in time, AI’s use outside white collar intellectual work will remain limited.
That means AI is perfect for summarization, collation, sifting vast amounts of data to glean insights. But it’s not actually very creative. It’s a pattern matcher. Want a version of PacMac with different graphics? Easy. Want to diagnose a rare disease? Hard. Devise a warp drive for space travel using known laws of physics but assembled in a way no one has ever seen before? Virtually impossible for AI.”
Have your own lesson to share? Email us at ai.newsletter@newsweek.com
Run Log
By Adam Mills
Clinical registries play a critical role in patient safety, but the work required to maintain them has long been dominated by manual abstraction. For one health system participating in the National Surgical Quality Improvement Program (NSQIP), that burden had reached a breaking point. Processing roughly 22,000 cases a year consumed more than 11,000 hours of clinician time, a workload Brent Dover, CEO of Carta Healthcare, described as “unsustainable.”
To address that strain, the system deployed Carta Healthcare’s Lighthouse platform, which Dover called a “Hybrid Intelligence platform fusing AI speed and accuracy with clinical expertise.”
Early adoption was cautious. “Seasoned abstractors were skeptical, double-checking every automated suggestion,” he said. Over time, confidence grew as the system surfaced details that even experienced reviewers might have missed. “Lighthouse doesn’t replace my judgment—it enhances it,” one abstractor told the team.
The operational impact was immediate. Average abstraction time per case was “cut nearly in half, saving up to 6,000 labor hours annually,” Dover said, while quality remained intact. The team consistently maintained “99 percent inter-rater reliability,” reinforcing that efficiency gains did not come at the expense of accuracy.
By shifting abstractors from “data hunters” to “data validators,” the system showed how AI can reduce administrative burden while keeping clinical judgment at the center of quality improvement.
Context Window
■ Cleveland Clinic researchers are using AI-driven analysis to help reverse Type 2 diabetes by identifying personalized care pathways, showing how machine learning could expand treatment beyond medication-heavy approaches like GLP-1s. [Newsweek]
■ South Korea will require all AI-generated advertising to be clearly labeled starting in 2026, a move intended to protect consumers from deceptive deepfake or fabricated content while balancing innovation and regulatory oversight. [AP News]
■ OpenAI rolled out GPT-5.2, its most capable model yet for professional work and long-running AI tasks, with the company saying it delivers stronger reasoning, better handling of lengthy inputs, as well as improved coding and productivity tools. [OpenAI]
■ Despite hype, many businesses still struggle to turn AI into profit and measurable value. Surveys from Forrester and BCG show only a small fraction of companies report clear margin gains, forcing a recalibration of expectations around adoption. [Reuters]
■ AWS CEO Matt Garman emphasized that AI should augment developers rather than replace them, and detailed AWS’s enterprise-focused AI strategy around customizable models and data integration. [Wired]
■ Automakers are using AI-powered simulations alongside traditional wind-tunnel testing to design cars that can handle extreme conditions, showing how digital tools are supporting, not replacing, real-world engineering. [Newsweek]
Transfer Protocol
Denise Dresser, formerly CEO of Slack and a senior Salesforce executive, has been named chief revenue officer at OpenAI, tasked with scaling global commercial operations and accelerating enterprise adoption of its AI models and tools.
Amin Vahdat, a longtime Google executive and vice president, has been elevated to chief technologist for AI infrastructure, leading efforts to scale the global compute backbone underpinning the company’s most advanced AI models.
Anthony Enzor-DeMeo, formerly a senior technology executive with experience leading digital platforms, has been named chief executive officer of Mozilla Corporation, bringing a renewed focus on responsible AI development alongside privacy-first product strategy.
Theodore “Ted” Tanner Jr., formerly chief technology and strategy officer at BigBear.ai, has been named chief technology officer at Leidos, accelerating the company’s deployment of AI, mission-critical software and cyber and quantum solutions as part of its technology leadership team.
Marshall Chapin, formerly a technology executive focused on advanced energy systems, has been named chief executive officer of GridAI, leading development of AI-driven grid-optimization and power-management platforms for hyperscale data-center operators.
Know someone on the move in AI? Send job change info to ai.newsletter@newsweek.com
Magic Moment
“My home last weekend, I literally got my new cameras up and running so I can watch who’s coming to my house…You know, license plate—no facial recognition. Like, there’s a car driving by, it can identify the license plate and then record it … I did everything. Yeah, it actually didn’t take that long. I can’t claim 100 percent credit for building every little bit of that code, but it was connecting the pieces… Everything I bought, every piece in there, every technology was completely built in the United States, to make sure my pictures aren’t flying somewhere outside of this country. It doesn’t even leave my network, so it stays completely on my network. It keeps my family safe. I like it, but it’s also just a way I could dabble and learn.”
Experience some AI magic? Tell us about it at ai.newsletter@newsweek.com
Tap or click here to get this newsletter delivered to your inbox.
