Another Crazy Day in AI: Building Trust in Resource Constrained Environments
- Wowza Team

- Sep 27
- 4 min read

Hello, AI Enthusiasts.
Made it to the weekend? Here’s something to think about as you log off.
AI is finding its place in classrooms and fields alike. Head of the Wadhwani Institute for AI in India shared how thoughtful tech can teach reading and support farmers —proof that impact comes from listening first.
If research has been piling up, a lead from Microsoft shares six ways their new AI agent can help you handle complex work without breaking a sweat.
And robots are stepping up too. Gemini Robotics 1.5 is learning to think before moving, making robots more helpful and less chaotic.
Step back from the week, relax, and enjoy a little tech-inspired wonder.
Here's another crazy day in AI:
Community centered approach to tech solutions
6 ways AI researcher makes your job easier
Next-gen robots learn, plan, and perform complex tasks
Some AI tools to try out
TODAY'S FEATURED ITEM: Practical Deployment of Social Impact Technology

Image Credit: Wowza (created with Ideogram)
What happens when you take artificial intelligence out of tech conferences and drop it into villages where internet is spotty and smartphones are shared?
In a recent episode of the Sidecar Sync Podcast, hosts Mallory Mejias and Amith Nagarajan speak with Shekar Sivasubramanian, Head of the Wadhwani Institute for Artificial Intelligence in India, about the practical use of AI to address social challenges. Shekar shares examples of AI applications designed to make a real difference, such as an oral reading fluency assistant supporting millions of children and adults across India, and smartphone-based pest detection tools helping rural farmers. The conversation explores the steps involved in deploying technology effectively in diverse and resource-constrained environments, including building trust with users and collaborating closely with local communities and government agencies.
What emerges from the discussion:
Working with available resources - Teams replace costly specialized equipment with items families already own, like wooden rulers for health measurements, adapting to what people actually have rather than what would be ideal
The bulk of work lies beyond coding - Sivasubramanian notes that AI algorithms account for roughly 3-5% of any project, while community engagement, training, and ongoing support consume the majority of resources and time
Proximity builds credibility - Staff members work directly within government ministries and live in target communities for months, establishing trust through consistent presence rather than brief visits
Economic realities at scale - Solutions must run at extremely low costs, around 5 paisa per use, to remain viable when serving millions of people while funding continuous improvements
Social networks extend reach - Agricultural tools work through "lead farmers" with smartphones who share information with neighbors lacking the technology, multiplying impact through existing relationships
Honest communication about failures - The institute openly discusses system limitations and error rates with users, incorporating human oversight rather than presenting AI as foolproof
Simplicity beats sophistication - Interfaces provide direct guidance instead of multiple options, recognizing that many users have limited experience with complex digital tools
Continuous learning from real use - Systems undergo regular updates based on new data and user feedback, with teams maintaining long-term relationships to track changing needs
This interview offers a window into the complexities that surface when technology moves beyond pilot programs toward actual deployment in challenging environments. Sivasubramanian recounts how one farmer told him, "I don't understand your technology, but I know you care," highlighting how acceptance often depends more on demonstrated commitment than technical prowess. The institute's work reveals gaps between laboratory success and field implementation that rarely get discussed in technology circles.
Working in environments with limited infrastructure, diverse languages, and varying literacy levels requires approaches that differ significantly from standard software development. While Wadhwani's projects show promise for addressing genuine needs at impressive scale, they also illustrate the substantial investment in relationships, cultural understanding, and patient iteration required to make sophisticated tools work in complex social settings. Their story provides a reality check on what it actually takes to deploy technology for social good, offering lessons that extend well beyond their specific context to anyone grappling with the gap between innovation and implementation.
Watch it on YouTube here.
Listen on Apple Podcasts here.
Listen on Spotify here.
OTHER INTERESTING AI HIGHLIGHTS:
6 Ways AI Researcher Makes Your Job Easier
/Vanessa Ho, Performance & Retail Ads Product Lead, APAC Go To Market at Google, on Microsoft - Source
Ever wish you had a research assistant that could do hours of work in minutes? Microsoft’s new AI agent, Researcher in 365 Copilot, does exactly that. It gathers and analyzes your emails, meeting notes, documents, and even external sources like news or industry blogs to produce in-depth reports. Unlike Copilot Chat, which gives quick answers, Researcher tackles complex, multi-step analysis, which is helpful for strategy planning, board presentations, customer insights, and more. It even shows its reasoning steps so you can verify results while staying within your organization’s data security policies.
Read more here.
Next-Gen Robots Learn, Plan, and Perform Complex Tasks
/Carolina Parada, Sr Director and Head of Robotics, on Google DeepMind Blogs
Robots are getting smarter, and Gemini Robotics 1.5 is leading the charge. This vision-language-action model can perceive, plan, and act on complex tasks, thinking through steps before moving. Paired with Gemini Robotics-ER 1.5 (which acts as the high-level “brain”), these models let robots reason, use tools, and generalize actions across different physical bodies. From sorting laundry by color to handling multi-step tasks safely, this duo allows robots to navigate the real world intelligently. With safety and alignment measures in place, these advancements are a key step toward general-purpose, physically capable AI.
Read more here.
SOME AI TOOLS TO TRY OUT:
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