In partnership with
Flow time: 10 min I your weekly pulse on AI news, tools and case studies reshaping the water sector
Find out why 100K+ engineers read The Code twice a week
Staying behind on tech trends can be a career killer.
But let’s face it, no one has hours to spare every week trying to stay updated.
That’s why over 100,000 engineers at companies like Google, Meta, and Apple read The Code twice a week.
Here’s why it works:
No fluff, just signal – Learn the most important tech news delivered in just two short emails.
Supercharge your skills – Get access to top research papers and resources that give you an edge in the industry.
See the future first – Discover what’s next before it hits the mainstream, so you can lead, not follow.
🔍 What’s in today’s flow
🔬 AI Research: AI predicts blooms, optimises treatment, and fills data gaps, but still struggles outside the lab
🤖 AI Scientist: $300 M startup aims to let AI run real experiments and discover new water materials
💧 Case Study: small utilities use AI to preserve operator know-how and automate daily tasks
🧠 Tool: Comet by Perplexity scores 16.5/20 - a fast, source-based browser for research
🕵️ AI Shadow: When AI leaks data or misfires, zero-trust and good governance keep it in check
🔧 Case study: AI Tackles the “Silver Tsunami” in Small Water Utilities
Source: aistudio.google.com
What happened
Across the U.S., small water utilities are facing a “silver tsunami” , a wave of retirements among experienced operators, leaving gaps in knowledge and staffing. Many of these utilities are turning to AI-powered systems to fill the void. Platforms using machine learning, predictive analytics, and virtual assistants are helping operators monitor systems, detect leaks, forecast maintenance, and automate compliance reporting, all without needing large technical teams.
How it’s being used
Knowledge capture: The Water Research Foundation’s “What Would Jerry Do?: Chlorine” project trained an AI chatbot on interviews with Jerry Kemp, an 82-year-old operator with over 40 years of experience. Operators can ask the AI questions and receive guidance , even “pro tips” in Jerry’s own words.
Staff training: DC Water created an internal AI tool that organizes videos of its experts explaining tasks, then generates quizzes and searchable content to support staff learning.
Data processing: LLMs can now read unstructured data, from handwritten notes and PDFs to video and audio, using optical character recognition (OCR) and computer vision. Some utilities even use AI to interpret camera footage in the field to identify leaks, rust, or whether a valve is open or closed.
Why it matters
For small utilities, AI isn’t replacing people, it’s amplifying their knowledge and capacity. These tools help preserve institutional memory, automate repetitive work, and give new operators access to decades of expertise.
Looking ahead, utilities will need to focus on data governance, training, and leadership buy-in to use these tools effectively. The message is clear: AI is no longer futuristic or out of reach, it’s a practical ally helping even the smallest utilities stay resilient as experienced operators retire.
🤖Latest in AI: Periodic Labs wants to build an AI scientist
Source: theaiinsider.tech
A new startup called Periodic Labs just came out of stealth with a massive $300 million seed round led by a16z , yep, the same people who backed OpenAI. The company’s goal? To build AI scientists who can actually run real experiments, not just analyze data.
They’re combining AI models with robotic labs, so the system can form hypotheses, run tests, and learn from the results all without waiting for humans to catch up. The team includes some of the biggest names in AI research, like people who worked on ChatGPT, DeepMind, and even the attention mechanism that powers modern AI.
Why it matters
This kind of tech could completely change how we discover and test new materials for water treatment, like better membranes, filters, or PFAS adsorbents. It could also help automate lab experiments, making it faster to test new coagulants or optimise treatment processes. In short, AI might soon be doing the boring lab work so water scientists can focus on the big idea
🔬AI research spotlight: AI in water quality
Source: iwaponline.com
A new review by researchers from the University of British Columbia and McGill University maps out how AI is reshaping water quality management, from real-time monitoring to contaminant forecasting. The study examines over a decade of applications across surface water, groundwater, and wastewater systems, highlighting where AI adds the most value and where major gaps remain.
Key insights
AI for early warning: machine learning models are now accurately predicting algal blooms, nutrient spikes, and pathogen risks using continuous sensor data.
Optimizing treatment: deep learning and hybrid models are being used to control coagulation, aeration, and disinfection processes, reducing chemical use and improving effluent quality.
Filling data gaps: AI can reconstruct missing water-quality data and detect faulty sensors, strengthening long-term datasets that utilities rely on.
Challenges: The authors caution that model transparency, data standardization, and field validation remain major hurdles. Many published models perform well in the lab but struggle in real-world utility conditions.
Why it matters
This research signals a shift from theory to practical AI deployment in water quality management. By using AI to interpret complex, multi-parameter datasets, utilities can move from reactive sampling to predictive, automated control. But the authors stress that for AI to deliver real value, utilities must invest in data governance, cross-disciplinary teams, and continuous model validation, not just algorithms.
🔧Trending tool: Comet
Source: perplexity.ai
Perplexity is an AI “answer engine” that searches the live web, cites sources, and now ships an AI-native browser called Comet that bakes research, summarisation, and workflow helpers right into your tabs.
Key features
Live web answers with citations and follow-up threads
Comet browser: page summaries, data extraction, side-panel chat; optional curated news via Comet Plus
Enterprise controls: SOC 2 Type II, GDPR/HIPAA, “no training on your enterprise data
⚖️ AI Tool Scorecard
Ease of use: ⭐⭐⭐⭐½ - dead simple query → cited answer; Comet feels natural for research (4.5/5)
Cost: ⭐⭐⭐⭐ - strong free tier; Pro/Enterprise pricing is reasonable; Plus is optional (4/5)
Security & privacy: ⭐⭐⭐⭐ - SOC 2 Type II; enterprise data not used for training; compliance badges(4/5)
Integration: ⭐⭐⭐⭐ - Works in-browser; file/Q&A flows; easy to slot into research + drafting (4/5)
Overall: 16.5/20 - A fast, source-anchored research companion; Comet’s free release makes it an easy default for water-sector desk research.
🕵️AI’s shadows: when AI turns against you
Source: aistudio.google.com
As artificial intelligence becomes embedded across workplaces, a new kind of insider risk is emerging, one created not by employees, but by the AI systems themselves. Generative AI chatbots, digital assistants, and autonomous agents now hold access once reserved for humans. When misused, compromised, or poorly governed, these tools can leak sensitive data, replicate malicious instructions, or act far beyond their intended scope.
A recent Verizon report found that nearly one in three breaches involved insider activity, and third-party breaches are on the rise. With AI systems integrated into email, cloud, and operational data, the boundaries of “insider” are blurring fast.
Why it matters
Utilities, councils, and infrastructure operators are increasingly using AI to optimise maintenance, analyse water quality data, and automate reporting. But when AI models are connected to control systems or internal files, weak governance can expose confidential designs, sensor data, or plant operations. A single prompt-injection attack or unmonitored chatbot can create a silent insider threat.
Takeaway
AI expands the definition of insider risk. To protect critical data and infrastructure, organisations should:
Apply zero-trust principles to all AI systems
Log and monitor AI prompts, data access, and actions
Enforce clear governance for approved tools and data use
Train teams on safe prompting and AI hygiene.
AI can be an ally, but without guardrails, it can also turn against you.
Thanks for reading! I hope you’ve enjoyed this week’s edition and look forward to seeing you next week!
Dr. Andrea G.T