Another Crazy Day in AI: When AI Gets Deployed at Global Scale
- Wowza Team
- Oct 15
- 5 min read
Hello, AI Enthusiasts.
How’s your week going so far? Feels like AI’s moving faster than the calendar again.
At TIME’s recent Impact Dinner, business and tech leaders explored what “AI transformation” actually looks like once the pilots end and the real work begins.
Meanwhile, Microsoft’s also putting AI through the ultimate stress test with a new cybersecurity benchmark.
And OpenAI’s setting the stage for even bigger systems that could push the boundaries of what’s possible.
Here's another crazy day in AI:
The ground-level view of AI in business
Microsoft introduces smarter benchmark for AI in cybersecurity
OpenAI expands infrastructure with Broadcom collaboration
Some AI tools to try out
TODAY'S FEATURED ITEM: The Reality of Enterprise AI Adoption
Image Credit: Wowza (created with Ideogram)
What does AI implementation actually look like when you have employees in nearly every country?
At a recent TIME Impact Dinner in San Francisco, three leaders gathered to discuss how artificial intelligence is unfolding in the business world. The panel featured Athina Kanioura, Chief Strategy and Transformation Officer at PepsiCo, Ravi Kumar S, CEO of Cognizant, and Jared Kaplan, Co-Founder and Chief Science Officer at Anthropic. Moderated by TIME's Executive Editor Nikhil Kumar and covered by journalist Zoe Corbyn, the conversation moved away from the usual focus on AI development to examine what's actually happening as these technologies get implemented in large organizations with hundreds of thousands of employees.
What came out of the conversation:
PepsiCo trained its entire workforce of 320,000 employees across all global markets in AI, treating the initiative as a board-level strategic priority with quarterly reviews and financial metrics
Cognizant is currently hiring more recent graduates than in previous periods, attributing this to AI tools that help newer employees build expertise faster
Research from Cognizant indicates that lower-performing employees see 36% productivity improvements with AI tools, while top performers see 17% gains
Deploying AI in enterprises involves different challenges than traditional software implementation, requiring ongoing supervision and a focus on behavioral outcomes rather than just technical deployment
PepsiCo published a formal responsible AI policy for public and investor review, and was the only non-technology company to submit responses to NIST's request for proposals on AI ethics
Anthropic supported California's SB 53 legislation, with the rationale that safer AI systems are more practical for enterprise integration and trust
The discussion addressed "context engineering" as an emerging priority, developing AI systems that understand specific organizational workflows rather than functioning generically
Applications in sustainable agriculture were highlighted, particularly for regenerative farming practices across global supply chains
Expectations include AI providing increasingly sophisticated professional and scientific guidance within the next year
While acknowledging near-term productivity benefits, panelists expressed uncertainty about longer-term economic impacts and stressed the importance of empirical research and careful policy development
The conversation surfaced questions that organizations are working through right now. Coordinating training across hundreds of thousands of people in different regions, speaking different languages, and doing vastly different types of work is a logistical puzzle that goes beyond technology adoption. Understanding how AI affects retail merchandisers, agricultural suppliers, software developers, and research scientists requires looking at each context separately. The three panelists brought their own institutional perspectives and priorities, and they didn't always see things the same way, especially around employment trends and what timelines are realistic. Those differences matter because they show how much is still being figured out, even by people directly involved in deploying these systems.
Employment came up repeatedly, but the discussion revealed how complicated the picture really is. Cognizant sees entry-level hiring increasing. Anthropic's leadership expressed concern about longer-term displacement in knowledge work. PepsiCo emphasized using AI to improve conditions for workers without college degrees. These aren't contradictory positions so much as different observations from different vantage points. What seems clear is that AI's impact varies significantly depending on what kind of work someone does, where they do it, and what resources their employer commits to training and transition support. The hopeful scenarios the panelists described—more opportunities for advancement, less time on repetitive tasks, better access to expertise—could certainly happen. Whether they happen broadly or narrowly is still an open question.
We're in a phase where major commitments are being made without complete information about where this leads. Companies are spending substantial money and political capital on AI initiatives. Researchers are trying to study impacts while they're unfolding. Regulators are writing rules for technology that keeps changing. The panelists seemed relatively comfortable operating in this environment of uncertainty, treating it as something to monitor closely rather than something that can be definitively predicted. That approach—attentive but not overconfident, engaged but aware of limitations—probably reflects the actual state of things more accurately than either uncritical enthusiasm or blanket skepticism would. The technology is here and being used at scale. What it ultimately means for work, for businesses, and for the people doing the work remains very much in motion.
Read the full article here.
Watch it on YouTube here.
OTHER INTERESTING AI HIGHLIGHTS:
Microsoft Introduces Smarter Benchmark
/Anand Mudgerikar, Senior Applied Machine Learning Engineer, on Microsoft
Microsoft has launched ExCyTIn-Bench, an open-source benchmark designed to evaluate how well AI systems handle real-world cybersecurity investigations. Unlike traditional tests that rely on trivia-style questions, it simulates realistic, multistage cyberattacks inside a secure Microsoft Azure environment. The tool helps CISOs and IT leaders measure how effectively AI can reason, adapt, and explain its findings in complex threat scenarios. Microsoft says it’s already using ExCyTIn-Bench internally to refine its AI security models and strengthen tools like Security Copilot and Defender.
Read more here.
OpenAI Expands Infrastructure with Broadcom Collaboration
/OpenAI
OpenAI and Broadcom have announced a multi-year collaboration to co-develop and deploy 10 gigawatts of custom AI accelerators by 2029. The new systems—built with OpenAI-designed chips and Broadcom’s Ethernet-based networking—aim to power the next generation of large-scale AI clusters. Leaders from both companies, including Sam Altman and Hock Tan, say the partnership will push the boundaries of AI infrastructure while improving power efficiency and scalability. OpenAI says this move brings them closer to building the infrastructure needed to support future breakthroughs in artificial general intelligence.
Read more here.
SOME AI TOOLS TO TRY OUT:
Modul – Make great presentations fast, no design skills needed.
Tasklet – Automate workflows with AI across thousands of integrations.
Tight Studio – Turn screen recordings into polished demos in minutes.
That’s a wrap on today’s Almost Daily craziness.
Catch us almost every day—almost! 😉
EXCITING NEWS:
The Another Crazy Day in AI newsletter is on LinkedIn!!!
Leveraging AI for Enhanced Content: As part of our commitment to exploring new technologies, we used AI to help curate and refine our newsletters. This enriches our content and keeps us at the forefront of digital innovation, ensuring you stay informed with the latest trends and developments.
Comments