Cornell AI Network Newsletter - September 2026

Welcome to your AI & ML digest

The Cornell AI Network connects Cornell graduates and affiliates working in AI and machine learning across industries and roles. We host talks, hands-on workshops, and member projects, and connect people through a global network. This is our monthly newsletter that highlights everything AI at the Cornell University.

Join our community: LinkedIn Company Page | the Cornell AI Network site

Announcements

Cornell AI ecosystem highlights

Cornell is leading a four-year, $7.5 million U.S. Department of Agriculture project to launch the Center of Excellence for Orchard Robotics, according to the Cornell Chronicle. CALS professor Manoj Karkee is heading the effort, which pairs academic and industry partners to build autonomous, vision-guided robots for labor-intensive orchard work such as thinning fruitlets and picking apples.
After a successful pilot last spring, Cornell's Center for Teaching Innovation and Cornell University Library are expanding the AI Critical Literacy Program to all incoming students, faculty, and staff this fall, according to the Cornell Chronicle. The four-module program runs on Canvas with hands-on instruction covering how generative AI works, its ethical implications, classroom use, and personal use policies.
Cornell's CAROW Yang-Tan Incubator has launched a research initiative, led by ILR School professor Sarosh Kuruvilla, on how artificial intelligence is changing specialized white-collar work, according to Cornell AI Initiative news. The project looks past manual-labor automation to study workflow disruption, workforce readiness, and ethics in fields such as law, journalism, design, and finance.

Upcoming Cornell-affiliated AI events for September 2026

Cornell Financial Engineering Manhattan 2026 Future of Finance & AI - Friday, September 11, 2026 | 8:30 AM to 5:30 PM ET | Cornell Tech, 2 West Loop Road, New York, NY | The forum brings together finance professionals and Cornellians to look at how artificial intelligence and quantitative modeling affect financial engineering.
Destination AI: Hospitality Summit 2026 - September 29 to 30, 2026 | Registration 12:45 PM and main stage 1:30 PM ET on the 29th, full day on the 30th | The National Housing Center, 1201 15th St NW, Washington, DC | Sold out, waitlist open. The summit covers AI-driven changes to customer experience and operations in hospitality, with alumni and industry leaders speaking.
2026 Cornell Quant Conference: 5th Annual - Tuesday, September 29, 2026 | 10:00 AM to 12:00 PM ET | Verizon Executive Education Center, Cornell Tech, New York, NY | Researchers, practitioners, and students meet on the Cornell Tech campus for sessions on algorithmic modeling, data science, and machine learning.
2026 Wall Street Quant Awards & AI Conference - Thursday, December 10, 2026 | 2:00 PM to 6:00 PM ET | 26 Liberty Street, New York, NY | The year-end event recognizes quantitative innovators and covers new AI technologies affecting Wall Street.

Upcoming featured event: ORIE Colloquium: Eric Balkanski (Columbia) (2026-09-08)

On Tuesday, September 8, 2026, Cornell's Operations Research and Information Engineering department will host Columbia University Associate Professor Eric Balkanski for an ORIE Colloquium in 253 Frank H. T. Rhodes Hall.
Balkanski will present "Online Selection via Linear Programming," on new results in online decision-making for knapsack secretaries, secretaries with predictions, and truthful secretaries.
The talk covers a two-step approach that uses order-statistics indistinguishable distributions to reduce cardinal problems to ordinal ones, then solves them with new linear program formulations.

Event details

Recent event recap: Human-Centered GenAI for Teaching and Learning (2026-09-03)

On September 3, 2026, Cornell's Center for Teaching Innovation hosted the in-person workshop "Human-Centered GenAI for Teaching and Learning" in the CCC 220 Multipurpose Room.
The session covered a human-centered framework and design principles to help instructors across disciplines build assignments for responsible, effective GenAI use.
Participants evaluated practical use cases and drafted their own AI-enhanced activities using Cornell-specific GenAI tools and resources available to faculty and students.

Event details

Full events calendar: cornellainetwork.com/events

Cornell AI startups & ventures

Cornell-connected companies in AI, curated from LinkedIn and Crunchbase.

Exostellar. This cloud and AI infrastructure tool automatically shifts applications across cloud servers to raise compute efficiency and cut operating costs. Qualcomm acquired the company to build into its AI product line, according to the Cornell Chronicle. Exostellar grew out of doctoral research by Zhiming Shen, Ph.D. '17, and was co-founded with Cornell Ann S. Bowers College of Computing and Information Science professors Hakim Weatherspoon and Robbert van Renesse, with incubation support from Cornell's Praxis Center for Venture Development and the Ignite gap funding program.
Makora. This GPU computing optimization suite, led by its flagship product MakoGenerate, is built to speed up GPU deployment and make AI workloads faster, more reliable, and cheaper to run. The company raised $8.5 million in seed funding to work on AI hardware and GPU scaling bottlenecks. Mohamed Abdelfattah, an assistant professor at Cornell Tech and Cornell Engineering, co-founded Makora with Waleed Atallah and Lukasz Dudziak.
MathGPT. This AI learning platform reads photos of STEM worksheets and generates step-by-step solutions, follow-up answers, practice tests, and custom video explanations. It has grown to 500,000 monthly active users and now has a 15-person team in San Francisco. Cornell Class of '26 computer science undergraduates Nour Gajial and Yanni Kouloumbis founded the company after an on-campus hackathon during their sophomore year.

Career opportunities

Featured AI/ML positions:

Founding Accountant at Porter - New York, NY - posted by Benjamin Lieblich, MBA '16
AI Engineer at FedEx - Dallas, TX and Memphis, TN - posted by Abiola Abraham, MBA '24
Sr Data and AI Specialist at Twilio - Remote - posted by Mustafa Ali, MEng '12
Postdoc AI at Weill Cornell Medicine - New York, NY - posted by Bobak Mosadegh, Professor

Full job board: cornellainetwork.com/jobs

The latest AI & ML news, by stack layer

Infrastructure layer

Cornell Bowers researchers are working on compute constraints through the Bowers-LinkedIn partnership. Doctoral student Ali Behrouz is designing next-generation transformer architectures with test-time memorization, and assistant professor Rachee Singh is building stateful tool-value caching to cut agent latency.

Data layer

Cornell Tech's Urban Tech Hub, working with New York City and New Yorkers for Parks, is feeding aerial imagery, urban tree census records, and 3D laser scans into The Generative Canopy Project to map block-by-block cooling capacity in heat-vulnerable neighborhoods, according to the Cornell Chronicle.

Model development layer

The Cornell University AI for Science Institute (CUAISci) is building domain-informed machine learning architectures for materials discovery, physics, and biology. The architectures tie model representations directly to the fundamental laws of nature.

Orchestration layer

The Cornell AI Innovation Hub is benchmarking LLM orchestration patterns, comparing structured Wiki-style LLM memory systems against conventional retrieval-augmented generation (RAG), to improve how context persists across complex institutional workflows.

Application interface layer

In emergency department settings, Cornell Bowers information science doctoral student Jonathan Segal is combining generative AI with augmented reality interfaces to flag clinical documentation errors and give physicians real-time guidance, according to Weill Cornell Medicine's AI program.

Cross-layer & industry

A two-day workshop in early August brought together 30 faculty and staff from across Cornell and Weill Cornell Medicine to test AI-driven software development. Many found the biggest challenges were not technical: defining goals, setting shared standards, and bringing the right people together mattered more than model capability.

STEEPLE: AI through seven lenses

STEEPLE = Social · Technological · Economic · Environmental · Political · Legal · Ethical

Social

How AI is changing human behavior, society, relationships, and access to information.

At the Digital Life Initiative (DLI), doctoral fellow Emma Harvey found that flawed crowdsourced data annotation undermines AI reliability and fair access to information. Studying corpora used to test whether chatbots give worse answers to prompts written in African American English, Harvey found that annotators introduced distortions, such as stripping out swear words unrelated to dialect. Her conclusion: machine learning benchmarks need measurement theory and real validity checks, not unvetted crowdsourcing.
The Cornell Initiative for Assistive Technology and AI (CATAI) is building assistive robotics, speech accessibility tools, and wireless sensing into everyday environments. Faculty including Thijs Roumen, Aditya Vashistha, Rajalakshmi Nandakumar, and Tapomayukh Bhattacharjee lead the effort, which is piloting deployments in New York City to support older adults and people with disabilities.
Helen Nissenbaum and Karen Levy continue researching how automated systems increasingly shape decisions that affect people's lives and privacy, through the AI, Policy, and Practice initiative and the Digital Life Initiative. Their work builds mechanisms to detect, explain, and correct fairness and privacy violations in automated decision-making systems before those violations widen social inequality.
A DLI essay by Travis Lloyd looks at how AI-generated content is changing behavior in online spaces. Lloyd argues that because generating text and media now takes much less effort, platforms face new problems in content filtering, media fragmentation, and how users judge whether digital content is genuine.

Technological

Breakthroughs, new architectures, tools, and technical capabilities.

Linxi Zhao and collaborators developed Limited Memory Language Models (LMLMs), a new approach to constrained-memory language representations that won a Best Paper Runner-Up award at the NeurIPS CCFM Workshop (Cornell NLP).
Graduate researcher Zoe introduced KnotGym, an interactive benchmark and manipulation framework recently covered by the Cornell Chronicle.
Cornell NLP reached two milestones in grounded natural language understanding. Alane Suhr won the AAAI/ACM SIGAI Dissertation Award for her work on interactive language reasoning, and Yoav Artzi won the TACL Test-of-Time Award for weakly supervised semantic parsers that map natural language instructions directly to executable actions.
Research teams at Robotics @ Cornell are advancing learning, autonomy, and multi-robot planning architectures for embodied perception and control across platforms including humanoids, aerial vehicles, and modular robots.

Economic

Market impacts, business transformation, job market shifts, investment trends.

Markets. AI appears to be attracting sustained capital rather than forming a uniform bubble. Cornell's Center for Data Science says the sector is "not in a bubble," even though some individual company valuations show signs of excess. That points to investors scrutinizing business fundamentals selectively, rather than pulling back across the board.
Business transformation. Cornell's AI Innovation Hub documents real deployments already running: a Treasury collaboration that recovered $100,000 in unidentified payments, and a TeamDynamix project that applies multi-agent AI to IT-support requests. Those examples, from the Cornell AI Innovation Hub blog, point to near-term adoption in finance operations, service management, and workflow automation, not just experimental pilots.
Work and skills. Cornell's August workshop brought together 30 faculty and staff from across Cornell and Weill Cornell Medicine to study AI-enabled software development, and the Innovation Hub highlights security, accessibility, documentation, and "skills" audits. The likely job-market effect is task reallocation: developers and technical staff increasingly need oversight, evaluation, security, and accessibility skills alongside coding.
Investment priorities. Cornell's data-science ecosystem is putting attention toward applied and responsible AI: climate monitoring, high-stakes decision-making, privacy, security, and fairness, rather than capability alone. Its Thought Summits offer $30,000 to $50,000 for interdisciplinary collaborations, a sign of continued institutional investment in research connecting AI methods to public and enterprise applications.
Strategic direction. Cornell's Center for Data Science for Enterprise and Society organizes its data science work around human-generated data, computational social science, digital agriculture, urban platforms, and the economics-computer-science interface. That agenda points to growing demand for AI investment tied to domain-specific data, governance, and measurable social or operational outcomes.

Environmental

AI's carbon footprint, climate applications, sustainability trade-offs.

Projects under the Cornell AI for Sustainability Initiative (CAISI) use sensor data and predictive modeling to adjust feed formulations and cut methane output in dairy and livestock systems. Related machine learning work is speeding up the search for stable, high-performance battery and fuel-cell materials.
Researchers at the Cornell University AI for Science Institute (CUAISci) are using deep learning to forecast how shifting climate patterns affect crop yields, pest distributions, and irrigation demand. They are also using multi-objective optimization algorithms to reduce conflict between renewable hydropower projects and aquaculture ecosystems.
The Institute for Computational Sustainability continues building data-driven frameworks to study biodiversity conservation, Arctic vegetation shifts, and environmental trade-offs in fragile ecosystems.
Cornell's NSF Research Traineeship (NRT) on AI for Sustainability trains graduate researchers across engineering and science disciplines to work on decarbonization, from molecular material discovery to energy-food-climate systems.

Political

Government policy, international competition, national AI strategies.

Cornell's Global AI Initiative frames international AI competition around inclusion and representation. It notes that many AI systems are built around WEIRD populations, who make up less than 15% of the world's population, and calls for global perspectives in how systems are designed, evaluated, and governed.
The initiative identifies governance, democracy, fairness, and environmental impact as strategic policy domains. Its focus areas include AI's effects on public discourse and civic participation, equitable deployment, the environmental costs of infrastructure, and the policy frameworks needed for transparency, accountability, and public oversight.
Cornell's AIPP program focuses on interdisciplinary AI governance rather than technology-only competition. Led by Solon Barocas, Jon Kleinberg, Karen Levy, and Helen Nissenbaum, it combines computational, legal, mathematical, philosophical, and sociological methods to bring together researchers, policymakers, and practitioners.

Legal

Copyright rulings, liability frameworks, data privacy law, regulation.

Copyright and the generative-AI supply chain. The Center for Law and AI's featured paper, Talkin' 'Bout AI Generation: Copyright and the Generative-AI Supply Chain, by Katherine Lee, A. Feder Cooper, and James Grimmelmann, maps where infringement questions arise at each stage from training data to generated output, and the choices facing legislators and regulators.
Liability, accountability, and legal practice. The Center studies how AI is changing legal practice and access to justice, and Cornell Law's curriculum now covers bias, fairness, transparency, and responsible AI use in legal work, according to Cornell Law School.

Ethical

Bias, fairness, consent, safety research, responsible AI development.

Cornell launched a required AI Critical Literacy Program for all incoming first-year undergraduates. The Generative AI Education Working Group and the Center for Teaching Innovation built the Canvas-based curriculum, which covers the ethical implications of foundation models, has students write personal AI use policies, and teaches how to evaluate model outputs critically, according to a Cornell Daily Sun column.
The Cornell ILR School's Center for Advanced Human Resource Studies (CAHRS) hosted a working group on how organizations use automated tools in talent acquisition. The session covered the ethical hurdles of candidate sourcing and evaluation, including algorithmic fairness, transparency, and early-career workforce expectations, with insights from ILR Director of Career Development Anu Lyons at a CAHRS session on early-career recruitment in the age of AI.
Analysis in The Cornell Daily Sun re-examined "AI doomerism" through the lens of historical labor protections, arguing that reducing harm and surveillance takes technical fluency, not avoidance. The piece argues that real accountability, such as the guardrails negotiated in recent collective bargaining agreements, comes from understanding system limitations and steering adoption policy directly.

This month's AI signals

Quick-scan digest: what to act on, watch, or guard against.

Green

With support from the Yang-Tan WorkABILITY Incubator at CAROW, ILR professors Sarosh Kuruvilla and Virginia Doellgast, along with Worker Institute senior research associate Zoë West, launched "AI and the New Informality." The two-year initiative is building practical tools, including an AI informalization diagnostic to assess how high-skill roles in fields such as finance, law, journalism, and design are losing formal protections, and a governance playbook with bargaining clauses and regulatory templates to preserve job quality, according to the ILR School announcement.
Led by Rene Kizilcec (Cornell Bowers CIS) and ILR professors JR Keller and Michèle Belot, the AI-Ready Workforce Initiative secured a $100,000 seed grant and an honorable mention from the Laude Institute. The team is building an AI-driven system that turns real workplace duties into interactive case simulations, moving hiring from static credentials toward measurable evaluations of candidate AI fluency, relational fluency, and adaptive flexibility.
A new project funded through CAROW's AI and Work Grant program is studying how generative models like ChatGPT interact with human decision-making biases during idea generation. The goal is to design AI systems that preserve idea diversity and prevent groups from converging on the same ideas during collaborative creative work.
Cornell Law School recruited five new faculty members, including Robert A. MacKenzie, director of the Entrepreneurship Law Clinic, adding to its clinical training and scholarship in artificial intelligence, civil procedure, and entrepreneurship law, per Cornell Law's announcement of the five hires.

Yellow

Cornell CALS researchers built the Biogeochemistry-Informed Neural Network (BINN), which runs 50 times faster than prior models at mapping soil organic carbon dynamics, according to the Cornell Chronicle.
Why watch with caution: Unlike standard generative AI that repurposes text, process-guided neural networks embed deterministic mechanistic models, such as the Community Land Model, directly into their architecture. The knowledge these systems infer stays tied to the underlying land models. Using them to guide planetary-scale climate interventions needs extensive ground-truth validation, so hardcoded model assumptions do not skew environmental policy.
Cornell researchers and LinkedIn have partnered on research into AI fairness and privacy-enhancing technologies.
Why watch with caution: When platforms try to put theoretical fairness and differential privacy into large-scale production matching and hiring algorithms, tension often shows up between utility, privacy loss, and disparate impact. Alumni in engineering leadership should watch how these algorithmic guarantees hold up against real-world feedback loops without creating new blind spots.
New AI frameworks are increasingly guiding carbon-offset initiatives, such as the global "4 per 1,000" soil target, as Michigan Farm News reported.
Why watch with caution: Machine learning applied across large regions has a history of spatial bias, favoring data-dense areas over under-sampled ones, plus parameter equifinality, where different parameter combinations produce the same prediction. If financial or regulatory carbon credits come to rely on AI-derived regional metrics, the models need rigorous uncertainty quantification, such as Monte Carlo dropout distributions, so agricultural and climate capital does not get misallocated.

Red

Agentic AI can turn model errors into real-world cyber harm. Cornell experts warn that connecting an unpredictable LLM to tools that run code, contact servers, or take other external actions creates risks well beyond a chatbot's bad response, including file deletion, service disruption, or unauthorized intrusion.
Cybersecurity is becoming a dual-use AI capability. As language models get better at finding vulnerabilities, they can help defenders harden systems, but they also lower the cost of offensive discovery and exploitation for attackers. Cornell's assessment is that broad access for defenders is essential if security is to improve instead of degrade.
School AI policies need developmental guardrails, not blanket tech enthusiasm. Cornell Tech's Daria Ng says AI use should be tied to clear learning goals and AI literacy. Child-development expert Kimberly Kopko cautions that for younger students, AI cannot replace teacher-guided learning, play, and sustained human relationships, as shared in a post on X.

AI term of the month

AI Critical Literacy

What it is and why it matters. AI critical literacy means understanding how artificial intelligence tools and large language models work, how they are built, and how to evaluate their output critically. As generative tools spread through professional and academic work, users need to handle ethical challenges such as bias, hallucinated output, and privacy risk, so AI tools support human judgment instead of undermining it.
At Cornell. Cornell expanded its AI Critical Literacy Program to all incoming students through four Canvas modules covering how generative tools work, personal use policies, and ethics vetted across disciplines. Faculty said building this literacy early matters for students headed into high-stakes fields, alongside broader biomedical and clinical AI work at Weill Cornell Medicine, where patient outcomes depend on responsible use of algorithmic recommendations.

Source: Wikipedia's Glossary of Artificial Intelligence

Research paper of the month

Generative AI in Research: Cornell Task Force Report

Cornell Task Force on Generative AI in Research | Cornell University, September 2026

The Cornell Task Force on Generative AI in Research, hosted by Cornell University, reviewed generative AI technologies across campus research and published institutional guidance on ethical use, reproducibility, and computational integrity.
The report sets out principles for faculty and student researchers, covering protocols for responsible experimentation, data privacy safeguards, provenance tracking, and safe deployment of frontier machine learning models in scientific workflows.
The initiative is meant as an ongoing framework for the university, with recommendations for institutional computing support, policy governance, and research infrastructure to keep academic rigor intact as autonomous and generative models evolve.

Read the paper →

Video of the month

Machine Learning Keynote

Cornell and invited experts | See the eCornell keynote library

Speakers: The keynote listing does not name individual speakers.
Date: The keynote listing does not give a specific date.
Description: The eCornell keynote lets Cornell AI alumni explore machine learning topics and applications on the university's digital learning platform. Alumni sign in with their NetID to stream the keynote and earn continuing education units from Cornell.

Watch on eCornell →

Community opportunities

Get involved

Weekly AI office hours. Every Thursday at 10 AM ET, open to Cornell AI project teams. Ask for the link at contact@cornellainetwork.com.

Open volunteer roles. Newsletter section editors and event coordinators. Introduce yourself on LinkedIn.

Cornell AI resources

Featured alumni spotlight

Cornell AI alumna Alane Suhr, Ph.D. '22, now on the faculty at UC Berkeley, received the AAAI/ACM SIGAI Dissertation Award for her doctoral dissertation, "Reasoning and Learning in Interactive Natural Language Systems," as noted by the Cornell NLP group.
Community impact. Suhr's award reflects continued momentum for the Cornell Natural Language Processing group, which connects researchers across the Ithaca campus and Cornell Tech in New York City, in computer science, information science, and linguistics.
Tell us your news. If you have published new work, started an academic role, or founded an AI venture, send your career update to the Cornell AI Alumni network to be featured in the next community spotlight.

About the Cornell AI Network

We are a community of Cornell graduates and affiliates working to advance AI and ML through collaboration, education, and ethical innovation. Our mission is to connect talent across industries, promote responsible AI development, and work for lasting impact on the field.

Connect with us: contact@cornellainetwork.com | LinkedIn | the Cornell AI Network site

Want to contribute content, suggest topics, or share opportunities? Reply to this newsletter or reach out to contact@cornellainetwork.com.