MIT Morningside Academy for Design

Term: September 01, 2023 - August 31, 2026

Project participants:

• Prof. Andres Sevtsuk, Department of Urban Studies and Planning at MIT
• Prof. Dr. Gerard de Melo, Chair of Artificial Intelligence and Intelligent Systems at HPI
• Dr. Rounaq Basu, Department of Urban Studies and Planning at MIT

Why do some streets attract more social activities than others? Is it the design of the street, or is it the land use pattern within neighborhoods that allows businesses around which people congregate and meet each other? Or is it instead the demographic characteristics of the residents that make some areas more prone to neighborly interactions than others? In a first-of-its-kind study, we will use big data to explore what facilitates social ties and face-to-face interactions on city streets, building stoops, and local businesses – the ‘sidewalk ballet’ – in New York City. This will involve developing state-of-the-art scalable computer vision models to detect how people interact with and within public spaces from open-source street view imagery. We will then create social activity and urban vibrancy indicators for every sidewalk in the pedestrian network. Finally, we will explore whether and how the intensity and diversity of social activities on city streets depend on characteristics of the streets themselves and the neighborhoods that surround them. The tools we create and publish will also allow similar analyses to be carried out for many other cities across the world, eventually leading to a global understanding of how vibrant urban spaces can be designed and supported.

Societal impact

Our project connects directly to SDG #11 (Sustainable Cities and Communities) by providing a better understanding of how to facilitate more sustainable travel (e.g., through walking and biking) through the design of vibrant urban neighborhoods. We also see implications for SDG #3 (Good Health and Well-Being) as both walking and spending quality time in urban spaces have been linked to better physical and mental health outcomes. Additionally, we anticipate that our exploration of the spatial variations in urban vibrancy can help address SDG #10 (Reduced Inequalities) by uncovering ways in which infrastructure investments (or the lack thereof) and land use patterns might adversely affect present-day opportunities for social interactions within historically marginalized communities.

Our project will create new knowledge about ways to design vibrant urban spaces that allow for stronger connections to communities and places as well as the opportunity for more sustainable travel.

Andres Sevtsuk

Term: September 01, 2023 - August 31, 2026

Project participants:

• Dr. Connor Coley, Department of Chemical Engineering, Department of Electrical Engineering and Computer Science at MIT, Henri Slezynger (1957) Career Development Assistant Professor
• Dr. Sumaiya Iqbal, Bioinformatics and Computational Biology, The Center for the Development of Therapeutics (CDoT), The Broad Institute of MIT and Harvard
• Prof. Dr. Bernhard Renard, Professor for Data Analytics and Computational Statistics at HPI
• Dr. Henrike O. Heyne, Senior Scientist for Genomic Medicine, Hasso-Plattner-Institute

Ion channels allow rapid transmission of information in living creatures, often across large distances. Due to their widespread importance, they have been targets of medication in humans (e.g. pain killers, antiarrhythmics) as well as insecticides, antibiotics or antimalarial drugs. In our project, we aim to combine information from protein biology and medical records to generate a publicly available resource that predicts the functional effects of mutations in ion channels using AI methods. Here, we will benefit from the functional similarity of different ion channel types that evolved from each other across many different species. We will map our results onto proteins from other organisms generating a publicly available resource of ion channel’s functional effects that should directly benefit precision medicine in humans as well as research in areas such as agriculture or infectious diseases. Our project will build on and strengthen previous successful collaborations. We will make our data and results available in a carefully designed web portal visualising scientific information, maximising convenience and user friendliness.

Societal impact

In our project, we aim to computationally predict functional effects of genetic variants in voltage-gated ion channels, map them to ion channels of other species and thus create a repository of ion channel variant effects for research communities in medicine, agriculture and infectious disease areas. We thus aim for contributing to United Nations Sustainable Development Goal 3 for health and well-being. Specifically, our prediction results should be directly usable to support treatment decisions of physicians treating individuals with rare diseases that are caused by genetic defects in ion channels. Our collaborating physicians are already using our prediction tool that we pioneered in sodium and calcium channels for that purpose. In addition, our results are already getting used in rare disease research which could eventually support drug development. In mapping our variant effect predictions onto ion channels in bacteria and human parasite species, our results have the potential to inform antibiotic and antimalarial research. As we plan to map (i.e., translate) our results onto ion channels in other species, our variant effects would be available for ion channels that are primary targets of insecticides. Thus, improved insecticides could support in agricultural productivity which is part of United Nations Sustainable Development Goal 2.

Our project aims to combine proteomics and transcriptomics data with clinical health records corresponding to the target gene and variants to generate a publicly available resource that predicts the functional effects of mutations in voltage-gated ion channels using AI methods.

Sumaiya Iqbal and Henrike Heyne

Term: September 01, 2023 - August 31, 2026

Project participants:

• Prof. Deblina Sarkar, Assistant Professor of Media Arts and Sciences at MIT, and AT&T Career Development Chair Professor at MIT Media Lab
• Prof. Dr. Ralf Herbrich, Managing Director and Head of the research group "Artificial Intelligence and Sustainability" at HPI

We are all witnessing AI revolutionize science, commerce, and the human experience. However, the AI of today employs large scale computation on traditional, von Neumann computers leading to massive energy consumptions. In fact, computers across the world are the fastest growing consumers of primary electricity and estimated to consume over 30% of world’s electricity by 2030. To allow sustainable consumption of the AI services, while ensuring judicious use of energy and minimal impact on the environment, we must transition to highly energy efficient paradigms of computation, encompassing algorithms and hardware. Contrary to von Neumann framework, the paradigm of neuromorphic computation features brain inspired principles of tight integration between processing and memory units, and approximate, error-tolerant results sufficient for meeting the final objective, enabling massive reductions in energy consumption and latency. Through our proposed project, we plan to leverage our collective expertise to holistically study and benchmark the performance of stochastic computing devices for AI applications. Our work would involve design, nanofabrication and characterization of such devices featuring controllable probabilistic switching behavior using 2D magnetic materials, theoretical and compact circuit modeling of the devices, and system level simulation for benchmarking their energy-bandwidth efficiency for AI applications against traditional computing devices.

Societal impact

Our project for enabling energy-efficient growth of AI aligns with several of the United Nations Sustainable Development goals, like SDG-13 climate action, SDG 12 sustainable consumption and by extension, SDG-7 affordable and clean energy. As indicated above, computation alone may account for up to 10% of the greenhouse gas emissions worldwide by the end of this decade. This can be avoided by transitioning to energy efficient computing paradigms as proposed in our work. By providing clear benchmarking of improvements in energy-efficiency and latency associated with our stochastic computing devices, we can accelerate this transition at scale, so that the services of large-scale computing like artificial intelligence can be consumed in an environmentally and practically sustainable manner. Finally, by capping the impeding energy demands from computation sector, we can ease the pressure on global energy supplies to help ensure affordable and clean energy.

Apart from global challenges, lack of adequate scaling in energy dissipation of computing hardware will start limiting the complexity of tasks that can be performed in mobile devices powered by batteries. Thus, there is an immediate need to design alternate, energy-efficient paradigms for computing, by reinventing in the material, device, architecture, and algorithm spaces. Through our proposed project, we intend to use our collective but complementary expertise across material science and electronic devices, and computing architectures and algorithms to provide a well-rounded solution to this impending problem.

Prof. Deblina Sarkar

Term: September 01, 2023 - August 31, 2026

Project participants:

• Jillian Kwong, PhD Research Scientist, Cybersecurity at MIT Sloan (CAMS)
• Keri E. Pearlson, DBA MIT Sloan School of Management; Executive Director, Cybersecurity at MIT Sloan (CAMS)
• Prof. Christian Doerr, Hasso Plattner Institute, Cybersecurity and Enterprise Security

As organizations shore up their cybersecurity defenses, small and medium small to medium-sized enterprises (SMEs) continue to lag behind. Although efforts are being mobilized across the public and private sector to address this gap, the question remains: how effective are these solutions and what can we learn from these new approaches and interventions? Early findings from the HPI-MIT seed funding round indicate existing mechanisms and solutions within the private sector primarily address symptoms of the problem by altering technical processes. However, core elements of the problem (i.e., mismatches in culture) are not being addressed and more emphasis needs to be placed on human processes like programs and training to embed and grow shared cybersecurity values across organizations. Building on ongoing HPI-MIT research, this study looks at how to develop a shared value model of cybersecurity that can be transmitted and adopted by organizations throughout the supply chain.

Societal impact

Cybersecurity is a prerequisite for many SDGs. For instance, body sensors monitor our health and doctors can make adjustments to implanted devices over the Internet. Many modern medical devices are essentially computers with medical sensors and telemedicine will be a key part in providing health care to all (SDG #3). The transition to renewable energy sources (SDG #7) requires working with distributed systems and energy grids that must be controlled remotely. Ensuring the cybersecurity of these systems is essential to avoid blackouts and other shutdowns of critical resources. Cyber security is essential for maintaining control of key industrial systems(#8), promoting the flow of goods, ideas, and knowledge around the world.

In fact, in a plethora of systems that we today find insecure and vulnerable, we find that security wasn’t frequently included or added as an afterthought. Today, engineers advocate ‘security by design,’ meaning that cybersecurity concerns should be considered and addressed during the conceptualization, design and manufacturing of an artifact.

Prof. Dr. Christian Dörr

Term: September 01, 2023 - August 31, 2026

Project participants:

• Prof. Dr. Martin Rinard, Department of Electrical Engineering and Computer Science at MIT and member of the Computer Science and Artificial Intelligence Laboratory
• Prof. Dr. Robert Hirschfeld, Software Architecture Group at HPI and the Digital Engineering Faculty at the University of Potsdam

Sustainable development demands effective coordination and management of resources. Software currently plays a central role in enabling this effective resource coordination and management. Designing such software typically requires close collaboration between domain experts and programmers to realize the capabilities required to achieve the desired impact. Our project seeks to bridge the gap between concrete needs expressed by domain experts and the abstract nature of code. The proposed approach to realizing this goal is to introduce interactive examples into the programming environment. Ideally, these examples will be meaningful to all stakeholders and will therefore comprise a strong foundation for the interdisciplinary communication required for the development of effective software. Our primary objectives are understanding the impact of live examples on software design, empowering stakeholders to create examples effectively, and designing and implementing example-based programming environments that promote design exploration and evolution and align with incremental and iterative design methodologies. A collaboration between Prof. Martin Rinard’s group at MIT and Prof. Robert Hirschfeld’s group at HPI is instrumental to automate example generation, reducing manual effort and promoting thorough exploration of design consequences.

Societal impact

Example-based programming and its potential align particularly well with the following SDGs:

• SDG #4 - Quality Education: A goal of the proposed research is to enable educators and developers to collaborate to deliver improved digital education pedagogy and software. Courses that involve or teach programming may benefit directly from examples that illustrate the operation, behavior, and intended functionality of the software. A methodology that prominently includes examples may promote more rapid understanding and contribute to digital literacy and IT skills. Also, programming as theory building is may be improved by tangible examples, which support the exploration and communication of domain models, consequences, and insights.
• SDG #9 - Industry, Innovation, and Infrastructure: Tangible, example-based programming tools may encourage efficient cross-disciplinary collaboration across all industries and helps to ensure needs are properly understood and implemented. Improved communication between stakeholders may reduce the chance of errors, which contributes to more resilient infrastructure. Enhanced documentation in the form of examples may ensure longer software life cycles through better maintainability, thereby saving resources needed to develop new software.

The proposed collaboration promises to reduce software designer and programmer effort via example automation and promote the exploration of the consequences of a design, specifically in potentially unanticipated regions of the input spaces to obtain a more complete understanding of a particular design and its implications.

Eva Krebs

Term: September 1, 2024 - August 31, 2026

Project participants: 

• Prof. Kripa Varanasi, MIT
• Prof. Christoph Lippert, Hasso Plattner Institute

This project addresses the critical problem of biomedical waste, specifically the over 300 million liters of cell culture waste generated annually. Conventional trypsinization, the primary method for cell dissociation, is labor intensive and generates significant waste, often resulting in cell damage and genetic mutations. Here, we propose a novel cell dissociation technology using electroactive materials with hierarchical microtextures for gentle cell detachment, complemented by a customized machine learning platform to analyze the spatio-temporal changes in cell morphology triggered by surface stimuli. This technology aims to reduce biomedical waste, protect sensitive primary cells and support high-throughput automated workflows. By minimizing waste and aligning with the United Nations' sustainability goals, our approach meets a critical need in cell culture and tissue engineering.

Term: September 1, 2024 - August 31, 2026

Project participants: 

• Prof. Marzyeh Ghassemi, MIT
• Prof. Ariel Stern, Hasso Plattner Institute

Despite the vast potential of digital health technologies, research indicates that a significant number of patient-users do not adhere to these technologies as intended. Low user retention prevents patients from fully benefiting from the applications and also undermines the sustainability of these technologies. Digital sustainability emphasizes maintaining the effectiveness of digital solutions within evolving contexts. For digital health technologies to be sustainable, we need to implement mechanisms to continuously learn from patients, and use this data to continuously improve the application, keeping the solution relevant and effective. Accordingly, this proposal aims to develop and test an Artificial Intelligence (AI) agent that remains updated on patient experiences by analyzing and enhancing text conversations in peer-to-peer support forums and capturing user needs. This project will explore two research questions (RQ): How can a human-AI collaboration enhance text-based conversations in peer-to-peer support networks (RQ1), and how effectively can it identify user needs arising from these interactions (RQ2)? By focusing on these areas, the project intends to address the challenge of low user retention in digital health technologies and contribute to the fields of digital health sustainability, AI in healthcare, and human-centered design. This will offer actionable insights for policymakers, and researchers.

Term: September 1, 2024 - August 31, 2026

Project participants: 

• Prof. Stefanie Mueller, MIT Computer Science and Artificial Intelligence Laboratory
• Prof. Patrick Baudisch, Hasso Plattner Institute, Research Group "Human Computer Interaction"

‘Design for Recycling' has focused on making products easier to recycle, but not on how lower-performance recycled materials affect design. We will explore opportunities in designing with recycled materials, particularly through methods like 3D printing, which allow for customized, one-off designs. We aim to create design tools supporting both manual product design and automatic product generation using Generative AI, tailored for manufacturing with recycled materials of varying structural properties.

Term: September 1, 2025 – August 31, 2026 

Project participants: 

• Prof. Caitlin Mueller, MIT, Architecture
• Prof. Gerard de Melo, HPI

Architectural structures are responsible for over 40% of global energy consumption and emissions, while also playing a crucial role in shaping human health and well-being. Integration of creative design with functionality and carbon efficiency is one of the greatest challenges in architecture today. To address this challenge, the team led by Professors Caitlin Mueller and Gerard de Melo is developing novel generative AI algorithms that link early-stage creative design exploration with high-fidelity engineering analysis and simulation.

Connecting widely used design modalities—such as images and text across varying levels of abstraction—to structural representations suitable for architectural and engineering simulation, this project aims to enable early environmental feedback and foster early-stage collaboration across disciplines. In doing so, the project advances the creation of low-carbon, high-quality architecture aligned with the United Nations Sustainable Development Goals.

Term: September 1, 2025 – August 31, 2026 

Project participants: 

• Prof. Connor Coley (MIT, Chemical Engineering)
• Prof. Dr. Bernard Renard (HPI)

Deep learning models have become powerful tools for predicting the functional impact of variants in proteins, with important applications in clinical genomics, drug development, and protein engineering. However, their lack of interpretability limits their adoption in domains where understanding model decisions is critical. This project aims to develop a concept-based interpretability framework tailored to protein fitness prediction. By linking model attributions to biologically meaningful concepts, such as structural sites and functional domains, the team seeks to create explanations that are both human-interpretable and biologically grounded.

Term: September 1, 2025 – August 31, 2026 

Project participants: 

• Prof. Pattie Maes (MIT, Media Lab)
• Prof. Dr. Falk Uebernickel (HPI) 

Led by Professors Pattie Maes and Falk Uebernickel, the team investigates real-time detection of problematic social media use (PSMU) through physiological and contextual signals. In order to address the growing psychological and behavioral consequences of compulsive digital engagement especially with short-form content, the project critiques the shortcomings of conventional interventions. Conventional interventions such as timers and screen time limits—which often overlook emotional state, user intention, and situational context.

This research aims to develop adaptive, real-time systems that move beyond static usage caps. By incorporating physiological data and contextual awareness, the project seeks to enable personalized and responsive interventions that support healthier, more sustainable relationships with technology.

Term: September 1, 2025 – August 31, 2026 

Project participants: 

• Professor Leonid Mirny (MIT, Physics)
• Prof. Dr. Helene Kretzmer (HPI)

The team led by Professors Helene Kretzmer and Leonid Mirny will build a deep learning model to explore the chromatin structure and epigenetic code of healthy and cancer cells. In some cancers, transcription factors—proteins that bind to DNA—can fuse together, giving rise to atypical DNA-binding sites and forming novel enhancers that drive disease-specific gene activity. The team will build and train an AI model to learn the differences between healthy and cancer cells and will leverage this model to design novel DNA sequences that are only active in specific cancer cells. These sequences will then be tested experimentally. The ultimate vision is that this work will form the basis for selective and sustainable cancer-specific therapies in the future.

Term: September 1, 2025 – August 31, 2026 

Project participants: 

• Prof. Scott Stern (MIT, Sloan)
• Prof. Dr. Ariel Dora Stern (HPI)

With the increasing usage of digital therapeutics (DTx), as digitally native tools, “real-world data” derived from outside of formal clinical trials (RWD) and “real-world evidence” (RWE) generated from such RWD are becoming essential resources for evaluating the effectiveness of new healthcare technologies. This project investigates how early-stage digital health companies—especially those developing DTx—strategically generate and communicate evidence to shape their market success and adoption.

Focusing on the intersection of entrepreneurial strategy and healthcare innovation, the project explores how firms navigate complex strategic environments and make decisions about what evidence to produce, how to disseminate it, and when to rely on collecting data through traditional clinical trials versus focusing on the use of RWD. With new policies, such as Germany’s upcoming RWE-linked reimbursement for DTx, these strategic choices are more relevant than ever. By comparing digital health firms with traditional pharmaceutical approaches, the project also draws broader lessons for evidence-based policy and innovation management.

Term: September 1, 2025 – August 31, 2026 

Project participants: 

• Prof. Antonio Torralba (MIT, EECS)
• Prof. Dr. Gerard de Melo (HPI) 

The United Nations has identified inclusive, equitable quality education as a key objective of the 2030 Agenda. Large language models and generative AI create unprecedented opportunities to advance this vision at scale. To maximize this potential and develop human-centered technologies for future sustainable instruction programs, we must integrate AI with proven educational practices that have fostered human learning for generations. Our approach centers on developing AI-powered educational models capable of engaging with education's most enduring and universal tool—the whiteboard —unlocking visual communication capabilities that current systems lack.

Using handwriting, sketching, and visual explanation in a shared visual space has long been integral to learning, enhancing comprehension, collaboration, creativity, and problem-solving.

However, existing AI-based tutoring systems remain limited to text-based interaction. While these methods can provide personalized feedback and answer questions, they cannot demonstrate mathematical proofs on a board or collaborate with students in constructing visual representations of complex ideas.

This project aims to develop the first whiteboard-literate AI tool for education that can explain concepts visually and interact with students in a shared visual thinking space. We strive to create transformative learning experiences, combining AI's personalization and extensive knowledge with the visual, embodied engagement that makes learning profound and lasting.

Term: September 1, 2025 – August 31, 2026 

Project participants: 

• Prof. Svafa Grönfeldt (MIT)
• Dr. Frank Pawlitschek (HPI) 

The research team led by Svafa Grönfeldt (MIT) and Frank Pawlitschek (HPI) investigates how academic institutions can serve as fertile ecosystems for digital ventures. As artificial intelligence becomes a central value proposition in emerging startups, the project explores how digital and spatial affordances within universities can be intentionally designed to enable sustainable venture creation and scaling.Through a series of empirical studies at innovation hubs such as HPI Engine and MITdesignX, the project maps out the role of university accelerators, digital infrastructures, and collaborative environments in fostering entrepreneurial success. By combining case study research, ecosystem analysis, and design science methods, the team aims to develop frameworks and models that guide institutions in building effective, hybrid (digital-physical) venture support systems. With a focus on translating academic excellence into scalable impact, Digital by Design offers strategic insights into how universities can evolve from knowledge centers into launchpads for high-impact, AI-enabled innovation.

Term: September 1, 2025 – August 31, 2026 

Project participants: 

• Prof. Miho Mazereeuw (MIT)
• Prof. Dr. Bernhard Renad (HPI) 

Team members from HPI and MIT explore automated and sustainable innovations to advance and broaden research in urban farming. Their work centers on 3D-printed hydroponic systems enhanced with solar panels, environmental sensors, and AI technologies to increase crop yield and minimize resources compared to conventional farming methods.

The project offers hands-on experiences through workshops, UX trials, and student-led research projects to collect data and perform statistical analyses—identifying where human or automation is best suited. By developing, testing, and iterating on hydroponic technologies, the team aims to push the boundaries of urban farming and contribute to the growing body of research in this field. Together, the teams will work with vendors and stakeholders to optimize hydroponic systems tailored for real-world performance and demonstrate their potential as sustainable solutions for the future of urban farming.

Term: March 01, 2023 - February 28, 2026

Project participants:

• Jillian Kwong, PhD Research Scientist, Cybersecurity at MIT Sloan (CAMS)
• Keri E. Pearlson, DBA MIT Sloan School of Management; Executive Director, Cybersecurity at MIT Sloan (CAMS)
• Prof. Christian Doerr, Hasso Plattner Institute, Research Group "Cybersecurity – Enterprise Security"

Compromised supply chains has become one of the biggest risks to businesses around the world. Not only do they threaten a company’s information assets but can also disrupt everyday operations. While companies of all sizes struggle with these issues, supply chain cybersecurity is especially difficult for small to medium-sized enterprises (SMEs). Despite the growing number of cybersecurity standards being passed around the world, research has shown some regulations can actually have an opposite effect by introducing greater levels of risk into cybersecurity supply chains. Rather than building resiliency, some government legislation has actually caused SMEs to leave the supply chain altogether (e.g., 2016 US Department of Defense (DoD) DFARS 252.204-7012 mandate). Our study explores how and why regulation fails to improve cyber-resiliency among SMEs and what can be done about it.

By understanding how companies implement new mandates and the barriers to improving SME cybersecurity, our project begins to shed light on why certain cyber-security regulations have the opposite of their intended effect. This research will allow us to provide recommendations for how SMEs can achieve higher levels of cyber-maturity as well as insights for regulators and larger organizations on how to support and drive change in SMEs.

Societal impact

Securing the cybersecurity supply chain is essential for enabling progress on all SDGs. As more activity, processes, and data move online into the cloud, it becomes essential that we focus on securing supply chains for organizations of all shapes and sizes. Enhancing security is crucial step in developing the basic infrastructure needed to carry out and make progress on any SDG. Specific goals it contributes to includes SDG #8.3 decent work and economic growth, as it supports SMEs, a pillar of the European economy, to cope with and remain competitive in view of increasing demands for cybersecurity regulation.

This research will provide much needed context for lawmakers, company leadership, and government officials as they think about and design new policy and improve regulation moving forward.

Keri Pearlson

Term: March 01, 2023 - August 31, 2025

Project participants:

• Prof. Regina Barzilay, Computer Science and Artificial Intelligence Laboratory, Abdul Latif Jameel Clinic for Machine Learning in Health, Massachusetts Institute of Technology
• Prof. Bernhard Renard, Hasso Plattner Institute, Data Analytics and Computational Statistics

Antibiotic resistance presents a serious threat to global health. We use deep learning with the goals of designing more sustainable, alternative antimicrobial agents for both clinical and agricultural use.  This can be achieved through the generation of new biomolecules, which can be used instead of traditional antibiotics to precisely target pathogenic microbes. We will use methods of explainable AI to gain insights into the biology of the designed molecules and thereby significantly advance the state of deep learning for synthetic biology.

Societal impact

Our study will not only significantly enhance the state of deep learning for synthetic biology, but also enable design of more sustainable, alternative antimicrobial agents for both clinical and agricultural use.  Therefore the project’s contributions will help ensure health and well-being (SDG #3) in the context of infectious diseases, while developing responsible consumption and production patterns for both antimicrobials and industries dependent on their use (SDG #12). Since more frequent emergence of pathogens due to climate change must be mitigated (SDG #13) innovation solutions in those two areas are urgently needed.

With the support of the HPI-MIT Designing for Sustainability Program, we will establish a close cooperation between the project partners to profit from our complementary competences.

Jakub Bartoszewicz

Term: September 01, 2023 - August 31, 2025

Project participants:

• Prof. John Fernandez, Professor of Building Technology in the Department of Architecture at MIT, Director, MIT Environmental Solutions Initiative, Founder and Director, MIT Urban Metabolism Group
• Prof. Dr. Gerard de Melo, Chair of Artificial Intelligence and Intelligent Systems at HPI
• Prof. Svafa Groenfeldt, Professor of Practice, School of Architecture and Planning at MIT, Founding member and Director of MITdesignX, Morningside Academy for Design
• Dr. Frank Pawlitschek, Director, School of Entrepreneurship at HPI, Managing Director HPI-SEED, Co-Founder ubitricity

Addressing global crises such as climate change requires not only expanding our knowledge but also designing methods and tools for timely action. Achieving planetary sustainability necessitates novel solutions from all sectors, particularly entrepreneurial innovation in the private sector. Startups, known for driving technological innovation and job creation, play a crucial role in this endeavor. However, the failure rate of startups is high, and their success is vital for society. To complicate matters, the startup ecosystem is characterized by uncertainty and volatility, compounded by growing uncertainties in climate and planetary systems. Therefore, there is an urgent need for a robust model that can objectively predict startup success and guide design for the Anthropocene. While startup success forecasting has gained popularity, existing research primarily focuses on venture capitalist selection processes rather than guiding the design of novel enterprises. This project aims to explore if AI-augmented decision-support systems can enhance current startup success forecasting systems by adding  explainability capabilities. The research proposal will utilize explainable AI and machine learning techniques to forecast potential startup success and provide reasoning for the forecasts. The study will train the according machine learning models based on the data of climate tech companies. The findings will be used to develop new tools for early-stage startup founders operating in complex and fast-evolving conditions, including specific environmental issues. This research has the potential to contribute to the design of sustainable and resilient startups, enabling effective action in addressing global crises.

Societal impact

Using the Nine Planetary Boundaries (Rockström, 2009) as a framework, the project will provide insights that entrepreneurs in the Global South do not currently have access to due to the lack of organized capacity in the form of local accelerators and incubators. In addition, the results will spur greater engagement of women entrepreneurs by providing guidance that is not dependent on local institutions and organizations that may perpetuate gender biases. Expanding the reach of expert guidance for entrepreneurs through the research results of this project, the SDGs that are directly addressed are – SDG-8 Decent Work and Economic Growth; SDG-9 Industry, Innovation and Infrastructure; SDG-10 Reduced Inequalities; SDG-5 Gender Equality; SDG-13 Climate Action.

This project aspires to explore the potential for AI-augmented decision-support-systems to extend and outperform current startup success forecasting systems by adding explainability capabilities. To accomplish that, we examine explainable AI / ML as a method to forecast potential startup success and provide reasoning for the forecasts. The unique advantage of utilizing ML techniques is the generalizability and adaptability of the model when perfect information is absent.

Svafa Grönfeldt (MIT) & Frank Pawlitschek (HPI)

Term: September 1, 2024 - August 31, 2025

Project participants: 

• Prof. Pattie Maes, MIT
• Prof. Falk Uebernickel, Hasso Plattner Institute

Dietary choices significantly contribute to obesity and carbon emissions. However, sustainable, future-oriented dietary decisions often require delaying immediate gratification. Prior studies show that episodic future thinking (EFT) interventions, which involve considering future self-related events, can enhance such long-term-oriented decision-making and help to reduce impulsive choices. We propose an innovative EFT intervention using virtual reality and generative AI to foster healthy, sustainable dietary choices by allowing users to interact with (i.e., see, talk to, and feel) their own AI-generated future selves. We will test its effectiveness in changing dietary decisions using well-established lab-based delay discounting and food choice paradigms, and examine the intervention’s psychological and neurophysiological mechanisms (e.g., attentional processes via eye tracking). This highly immersive EFT approach aims to promote future-oriented dietary choices aligned with health and environmental goals, contributing to SDGs 3, 12, and 13.

Term: March 01, 2023 - August 31, 2024

Project participants:

• Prof. Stefanie Mueller, MIT Computer Science and Artificial Intelligence Laboratory
• Prof. Patrick Baudisch, Hasso Plattner Institute, Research Group "Human Computer Interaction

The laser cutting design software Kyub, developed in Patrick Baudisch’ lab at Hasso Plattner Institute, allows users to design and fabricate high-quality physical prototypes with the help of subtractive fabrication machines, such as laser cutters. In this project the team of researchers from MIT and HPI will introduce the notion of sustainability into the software. The team will develop software that will suggest, during editing and in real-time, how to reduce material consumption, while minimizing impact in the design intent. This will build on a fast layout algorithm developed at Stefanie Mueller’s lab at MIT. The team will then evaluate how well both objectives, i.e., design intent and material efficiency can be achieved at the same time.

Societal impact

The work focuses on the sustainable development goal #12 Responsible Consumption and Production (reducing material waste by optimizing material usage), #9 Industry, Innovation and Infrastructure (by developing novel digital manufacturing strategies), and #4 Quality Education (by teaching high school students about novel digital manufacturing processes, such as laser cutting, and the need to be mindful of material usage).

In the digital realm, personalization is everywhere, and thus we can assume that with the rise of digital manufacturing and the option to personalize physical products, there will be a similar large demand. It therefore is crucial to consider the impact of personalization of physical products on sustainability now and we do this in our work by investigating the trade-off between personalizing physical goods and the impact on material usage.

Stefanie Mueller (MIT)

Term: September 01, 2023 - August 31, 2024

Project participants:

• Prof. Hiroshi Ishii, Jerome B. Wiesner Professor of Media Arts and Sciences at MIT, Associate Director, MIT Media Lab, Head, Tangible Media Group (Team Members: Dr. JB Labrune, Cyrus Clarke, Lucy Li)
• Prof. Dr. Bert Arnrich, Head of the research group "Digital Health - Connected Healthcare" at HPI (Team Members: Dr. Julia von Thienen, Corinna Jaschek, Luca Hilbrich, Philipp Steigerwald, Tim Strauch, Holly McKee

To create sustainable designs, it is crucial for designers to consider the broader impact of their solutions, not only on selected users, but on entire ecosystems. This necessitates transitioning from an egocentric to an ecocentric work approach, where the evaluation of design impact extends beyond human user needs and encompasses the effects on human-nature relationships as well. However, there are methodological challenges, such as comprehending the consequences of designs not only for specific user groups, but also addressing a great diversity of needs within ecosystems. This proposal aims to address these challenges by creating tangible perspective-taking objects to help designers empathize with varying needs in different ecosystems, integrating real-time information and supporting sustainable design decisions. The project also develops a comprehensive observation and analysis framework to assess sustainability in design, which will be rendered available beyond this research group to facilitate assessments in diverse projects within the HPI-MIT program.

Societal impact

This project makes a significant contribution by developing a comprehensive set of measurement tools, encompassing both automated and expert-based methods, to effectively track the sustainability of design practices. These measurement tools involve assessing the level of attention dedicated by design teams to each of the 17 rubrics of sustainability outlined in the UN Sustainable Development Goals and the Ecosystem Services & Biodiversity (ESB) framework of the FAO. Additionally, metrics will be devised to evaluate the sustainability of design outcomes, both in broader terms and specifically concerning their anticipated impacts on distinct ecosystems.

The project team co-hosted a workshop at the Berlin Science Week this November: https://berlinscienceweek.com/programme/waste-wearables-speculative-design-workshop-more-human-sensing

Our approach involves testing the impact of perspective-taking objects on the designer’s mindset, and on the sustainability of emerging design solutions in creative sessions.

Telesymbiosis Team