Cross-functional Teams
Highlights
Top Insights
In a new experimental study, professionals worked in the early stages of new product development with or without a GPT-4-based AI tool, and either working alone or with another professional. Those working alone with AI matched the performance of teams without AI. Without AI, R&D professionals tended to suggest more technical solutions, while business professionals tended to develop commercially oriented solutions. Using AI, all these professionals produced balanced solutions. Working with AI led to higher levels of self-reported positive emotions. Individuals using AI report positive emotional responses that match or exceed those working in a team without AI. These results suggest that AI can replicate some benefits of human collaboration.
Source: MIT Sloan Management Review; Working Paper
Top News
1. OpenAI released o1-pro, its most expensive reasoning model, promising more reliable answers.
2. Baidu has launched the reasoning-focused ERNIE X1, the first deep thinking model that uses tools autonomously, and the advanced multimodal ERNIE 4.5.
3. Roblox has launched and open-sourced Cube 3D, a foundational generative AI model that creates fully functional 3D objects and scenes from text prompts.
4. Nvidia unveiled the GR00T N1 foundational model for humanoid robots, and its open Llama Nemotron reasoning AI models for agentic AI platforms.
5. BYD has unveiled a new EV platform capable of 1,000 kW supercharging—enabling a 400 km range in just 5 minutes.
Additional Insights
1. AI Cannot Respectfully Evaluate Employees (California Management Review)
While AI is increasingly used in recruitment and employee evaluations for efficiency and fairness, recent research shows that it often undermines a critical workplace value: respect. Across multiple studies, individuals consistently felt less respected when assessed by algorithms rather than humans—even when bias was controlled for—because they believe only humans can fully understand their unique qualities. This perceived lack of dignity not only affects the individuals evaluated but also observers, potentially harming organizational culture and performance. To mitigate this, companies should integrate human evaluators into AI-driven processes, clearly communicate AI’s role, train managers on respectful evaluation, and adopt ethical standards for AI use in HR. Ultimately, ensuring respectful treatment is as vital as reducing bias in deploying AI for employee assessment.
2. Strategic alliances for gen AI: How to build them and make them work (McKinsey)
To unlock the full value of generative AI (GenAI), companies must move beyond traditional vendor relationships and form strategic alliances built on trust, transparency, and collaboration. Successful alliances require joint planning, cocreation of solutions, and shared investment and risk, particularly as GenAI remains a rapidly evolving, complex, and resource-intensive technology. Key criteria for selecting providers include scalability, interoperability, and reusability, with an emphasis on aligning incentives through outcome-based compensation and clear intellectual property boundaries. Organizations should also maintain control by building flexible infrastructure, continuously monitoring model performance, and assigning dedicated relationship managers. Ultimately, these partnerships must be deeply integrated, with governance structures and alliance strategies that evolve alongside the technology to achieve sustainable, enterprise-scale GenAI success.
3. Dreams of improving the human race are no longer science fiction (The Economist)
The movement to enhance human capabilities through technology, drugs, and genetic engineering is gaining momentum, driven by billionaire investors like Christian Angermayer, Peter Thiel, Sam Altman, and Bryan Johnson. Their visions range from reversing aging and improving mental performance to allowing athletes to compete while enhanced, as in the proposed “Enhanced Games.” Despite limited clinical evidence and ethical concerns, the industry spans supplements, gene therapies, and brain-computer interfaces (BCIs), with ventures like Neuralink and Rejuvenate Bio pushing boundaries. While regulators remain cautious, especially in the U.S., signs of a more permissive stance under political figures aligned with the MAGA movement suggest this could change. Public opinion remains divided, with support for medical restoration but skepticism toward augmentations beyond normal human limits—yet proponents argue a transformation toward “superhumanity” is not only possible, but inevitable.
Innovation Radar
1. AI Model Releases and Advancements
OpenAI has released o1-pro, its most expensive and compute-intensive reasoning model yet, promising more reliable answers despite only modest performance gains over the standard o1 (TechCrunch). OpenAI has launched next-generation speech-to-text and text-to-speech models in its API, enabling developers to build more accurate, customizable voice agents with natural, expressive interactions across diverse languages and scenarios (OpenAI).
Baidu has launched two new AI models, including the reasoning-focused ERNIE X1, the first deep thinking model that uses tools autonomously, and the advanced multimodal ERNIE 4.5 (Reuters).
Roblox has launched and open-sourced Cube 3D, a foundational generative AI model that creates fully functional 3D (and eventually 4D) objects and scenes from text prompts, aiming to accelerate and democratize immersive content creation both on and beyond its platform (Roblox).
Mistral has released Mistral Small 3.1, an open-source, lightweight, and high-performance AI model with improved text and multimodal capabilities, a 128k token context window, and superior benchmarks across multiple tasks—positioning it as the leading small model in its class (Mistral).
At its 2025 GTC conference, Nvidia unveiled a suite of AI-focused innovations—including the Blackwell Ultra GPU, Vera Rubin and Feynman chip architectures, DGX personal AI computers, advanced networking chips, free Dynamo reasoning software, and the GR00T N1 foundational model for humanoid robots—reinforcing its leadership in AI computing (Reuters).
NVIDIA has launched its open Llama Nemotron reasoning AI models to power agentic AI platforms, enabling developers and enterprises to build advanced, business-ready AI agents with enhanced reasoning, speed, and collaboration capabilities (NVIDIA).
Stability AI has released Stable Virtual Camera, a new AI model that transforms 2D images into immersive 3D-like videos with customizable camera paths and perspectives, available for noncommercial research use (TechCrunch).
2. AI Tools and Features
Anthropic added web search capabilities to Claude (Claude).
Vogent has launched beta self-learning voice agents that autonomously design, train, and improve themselves using reinforcement learning from unstructured data and real-time call feedback, enabling high-accuracy performance in critical industries (Vogent).
Patronus AI has launched the first Multimodal LLM-as-a-Judge, powered by Google Gemini, to evaluate and optimize image-to-text AI systems with advanced tools like Judge-Image, already adopted by companies like Etsy to enhance caption accuracy and reduce hallucinations (PR Newswire).
Google’s Gemini now offers two new features—Canvas, an interactive space for real-time document and code collaboration, and Audio Overview, which turns files into podcast-style summaries—enhancing creativity, productivity, and learning for users (Google).
Google’s NotebookLM now has an Interactive Mind Map feature (Google).
At Adobe Summit 2025, Adobe unveiled its unified AI platform and new agentic AI innovations—like the Experience Platform Agent Orchestrator and GenStudio updates—to power scalable, personalized customer experiences by integrating creativity, marketing, and data across its ecosystem (Adobe).
3. AI in Consumer Products
Tern AI has developed a low-cost, GPS-free vehicle positioning system that uses only map data and existing car sensors, offering a scalable, private, and resilient alternative to traditional satellite-based navigation (TechCrunch).
Zoom has unveiled major updates to its AI Companion, introducing over 45 new agentic AI skills and specialized agents across its platform to automate tasks, enhance productivity, and improve customer experiences through personalized, multimodal, and customizable AI-driven interactions (Zoom).
Muse has launched Muse S Athena, the first consumer wearable combining EEG and fNIRS technologies to deliver real-time, AI-powered cognitive fitness insights and personalized brain training for focus, resilience, and long-term brain health (Business Wire).
4. AI for Research and Social Impact
At its annual “The Check Up” event, Google announced six new AI-powered health initiatives—including enhanced health search results, global medical records APIs, pulse detection on Pixel Watch 3, an AI co-scientist for research, TxGemma for drug discovery, and Capricorn for pediatric cancer treatment—demonstrating AI’s growing role in improving global health outcomes (Google).
5. Other
a. Quantum computing
The launch of the Jinan-1 quantum microsatellite marks a major step toward practical, secure global quantum communication, enabling real-time quantum key distribution with compact, portable ground stations and laying the groundwork for scalable satellite-based quantum networks (Nature).
b. EV
BYD has unveiled a new EV platform capable of 1,000 kW supercharging—enabling a 400 km range in just 5 minutes—and announced plans to build over 4,000 ultra-fast charging stations across China to support the technology (Reuters)
