Quick Hits in AI News: New Research from Stanford and Others

Each week, as SHRM’s executive in residence for AI+HI, I scour the media landscape to bring you expert summaries of the biggest AI headlines — and what they mean for you and your business.

1. The 2025 AI Index Report: 10 Trends Defining the State of AI 

The Stanford Institute for Human-Centered AI recently released The 2025 AI Index Report. If you want the clearest snapshot of where artificial intelligence stands today — and where it’s headed — this is the report to read.  

What to Know:

1. Smaller models perform better: Phi-3-mini (3.8B params) matched performance once requiring 540B params.
2. Costs dropped massively: Querying GPT-3.5-level models is now 280 times cheaper than in 2022.
3. China closes the gap: The U.S. still leads in models, but Chinese models nearly match on benchmarks.
4. AI incidents surged: Harmful use cases rose 56%, hitting a record 233 incidents in 2024.
5. Agents show early promise: AI agents outperform humans in short tasks but lag in longer ones.
6. Investment soared: U.S. AI investment hit $109 billion; generative AI (GenAI) alone drew $33.9 billion globally.
7. Corporate AI use jumped: 78% of the firms surveyed use AI, up from 55% in 2023.
8. Health AI exploded: The FDA approved 223 AI-enabled medical devices in 2023.
9. State-level regulation is rising: U.S. states passed 131 AI-related bills in 2024.
10. AI optimism varies globally: China, Indonesia, and Thailand lead, while the U.S. remains skeptical.

Why It Matters: 

The report is Stanford HAI’s comprehensive, data-driven overview of how AI is evolving across research, economics, society, and governance. The research equips leaders with trends that matter — beyond the hype — showing where AI is advancing; where it’s failing; and how it’s reshaping business, policy, and daily life. Use it as a benchmark to assess your own organization’s readiness for a fast-accelerating AI future.  

Here are my main takeaways:

• AI continues to match — and, in some cases, surpass — human performance, even on the most demanding benchmarks. In 2023, new tests pushed the limits of advanced AI systems. Just a year later, scores surged — up 18.8, 48.9, and 67.3 percentage points on college-level multimodal reasoning (MMMU), grad-level sciences (GPQA), and advanced coding (SWE-bench), respectively. Beyond benchmarks, AI made major leaps in video generation, and language model agents began outperforming humans in time-limited coding tasks.
• AI is embedding into everyday life. Taking root in industries such as health care and transport, AI is moving from lab to real-world use. In 2023, the FDA approved 223 AI-enabled medical devices (up from six in 2015). On roads, self-driving cars are no longer experimental: Waymo now runs more than 150,000 autonomous rides weekly, and Baidu’s Apollo Go serves cities across China.
• Business is going all-in on AI. U.S. private AI investment hit $109.1 billion in 2024 — 12 times China’s $9.3 billion and 24 times the UK’s $4.5 billion. GenAI led the charge, drawing $33.9 billion globally (up 18.7% year over year). AI adoption also surged: 78% of organizations used AI in 2024, up from 55% in 2023. Studies confirm productivity gains and narrowed skills gaps.

2. Being Human in 2035: How AI May Transform Human Thought, Emotion, and Behavior

What to Know: 

A new report from Elon University’s Imagining the Digital Future Center surveyed nearly 300 global experts on how AI will affect human nature by 2035. Most said the changes would be “deep and meaningful” or even “revolutionary.” Nearly 200 experts also wrote essays predicting shifts in how people think, feel, act, and relate.  

Concerns focused on the negative impacts of AI on mental well-being, critical thinking, trust in social norms, self-identity, and empathy. The report explores 12 key areas of human behavior, including decision-making, creativity, and agency​.

Why It Matters: 

This report offers a rare expert view not just of AI’s technical or economic impact, but of how AI may reshape the human experience itself. For HR, learning and development, and policy leaders, the report also highlights the urgency of designing interventions that preserve human strengths such as empathy, curiosity, and purpose as AI tools become ubiquitous.

3. The Future of Collaboration: How AI Is Supplementing Teamwork and Innovation at P&G

What to Know: 

P&G partnered with Harvard and Wharton to study AI’s effect on teamwork through a live hackathon. Teams using AI worked 12% faster and delivered more balanced, cross-functional solutions. AI also boosted morale, with employees responding more positively when collaborating with AI tools. P&G sees AI not just as a tool but as a teammate that enhances innovation speed and creative output.

Why It Matters: 

For HR and business leaders, this highlights AI’s ability to improve collaboration, employee experience, and innovation outcomes. Success depends on clear integration strategies, training, and ongoing experimentation. P&G continues to invest in upskilling employees and embedding AI across product development, operations, and culture.

4. From AI Agent Hype to Practicality: Why Fit Matters More Than Flash

What to Know: 

AI agents are being overhyped, according to VentureBeat, but most enterprises fail to deploy them effectively. The authors argue that companies must focus less on flashy capabilities and more on where AI adds real value. Many leaders only automate current tasks, missing broader potential.  

A better strategy starts with mapping total value creation, then identifying new areas AI can help expand — not just speed up — work. The SPAR framework (Sense, Plan, Act, Reflect) is offered to help understand and design more capable agents.

Why It Matters: 

Most AI rollouts fail because they focus narrowly on automating tasks rather than reimagining work. HR and business leaders should shift from use-case checklists to value-based planning. By focusing on fit, not hype, teams can use AI agents to create new value, improve workflows, and unlock long-term growth. This requires strategic planning, experimentation, and the rethinking of roles and outcomes alongside technology.