LWL #40 Engineering the Future? AI Laboratories Around the World

Anthony Deen, Evelin Lasarga, Guillermo Romero, Ivette Yáñez Soria, Sara Ortiz Sep 08 2022 Blog


Engineering the Future ? AI Laboratories Around the World

It seems that we are witnessing significant progress in the development of AI systems and technologies almost daily. Behind the scenes, there are many dedicated organizations, laboratories, and research centers that are working arduously towards finding solutions to pressing issues. There are the well-known big tech giants, whose labs include Microsoft AI Lab, Google AI, OpenAI, DeepMind,  Baidu Research, Tencent AI Lab, that are renowned for focusing their efforts towards developing not only AI products that appeal directly to us as customers, but also on research that digs deep into improving the functionality of AI in general. These companies spend massive amounts of resources on research and development to create these new technologies. However, there are many more AI research laboratories across the world that are also contributing to investigating new forms, improvements, and ethical guidelines related to AI. A key actor in this endeavor is the academic sector, which is receiving more funding to focus on enhancing AI. Top universities across the world have also started research centers which contribute (through publications and research) to a better understanding of AI technologies. These include Stanford Artificial Intelligence Laboratory, MIT- Watson AI Lab, Artificial Intelligence Laboratory University of Tsukuba, Makerere Artificial Intelligence Lab , UMA AI Lab, UNAM Laboratorio de Inteligencia Artificial y Alta Tecnología. Other universities have also created alliances and laboratories, such as the Alan Turing Institute.  Given the need for continued research and knowledge production on AI, many independent laboratories have also sprung up and are currently contributing to the field through their work. These laboratories are normally composed of a mix of actors from different fields and sectors that have gathered to bring their expertise. These labs include AILab.One, Midjourney, IA Latam, NAIXUS, AI21Labs and many others. 

What is AI and why is AI Research Important? 

Why are so many institutions invested in developing research for AI? Why has it become so important? In 1950 Alan Turing asked a very important question that would expand the limits in which humans would utilize machines, which was: Can machines think?”. Based on this simple inquiry, efforts towards creating more advanced and complex machines and systems began. To define it simply, Artificial Intelligence is a field in which computer science, machines, and robust datasets are leveraged to enable problem solving systems that mimic the human mind. At its core, AI systems are fed information that allows them to perceive environments, make decisions, solve complex problems, imitate patterns, and recognize objects. Using these methods and techniques like Machine Learning (ML) and Deep Learning (DL), AI systems can facilitate innovative solutions to various real-world problems. 

The importance of AI lies in its intrinsic characteristics, which allows it to constantly learn on the large amounts of data it is given, and create automated solutions or responses to different problems, including in fields such as healthcare, finance, agriculture and even art. Moreover, AI has proven to facilitate and automate many daily tasks humans do and has therefore become an invaluable tool in streamlining many common operations. For this reason, we are seeing an increasing number of companies, academic institutions, government agencies and other sectors getting involved in researching and applying AI in their work. However, AI is not a silver bullet to all of humanity’s problems. AI systems raise many ethical considerations and issues with data bias, security, and privacy that affect how the systems perform are also relevant. This is where the importance of research in AI comes in. The research centers focus not only on how to create more innovative AI systems, but also in improving the existing systems and accounting for ethical issues, such as data-set biases and privacy concerns. In this research field, the work of independent researchers and research centers is extremely important, particularly when big tech companies do not seem committed to enacting principles of Ethical AI. Therefore, it is crucial to continue on the quest of researching new and different AI systems, their ability to solve problems, and ways they can be improved in accordance with ethical principles. 

Join us in this 40th anniversary version of Links We Like, as we discover different AI research laboratories across the world and the work they are doing to create new AI systems and conduct research to improve those that already exist. 


Established in 2011, the Centre for Artificial Intelligence Research is an innovative research network whose objective is to build AI research capacity in South Africa, and with it facilitate broader access to AI technologies and tools in the country. In the Centre, master and doctoral students across six universities in AI AND are trained to produce research through nine established research groups: Adaptive and Cognitive Systems, AI and Cybersecurity, AI for Development, Applications of Machine Learning, Computational Logic, Ethics of AI, Foundations of Machine Learning, Knowledge Representation and Reasoning, and Probabilistic Modeling. CAIR follows a hub-and-spoke model that releases periodic publications on topics related to their research group issues, and leads events such as the annual Southern African Conference for Artificial Intelligence Research.

Meta AI is the outgrowth of the Facebook Artificial Intelligence Research (FAIR) lab, which was the first of its kind within the organization. The current iteration of the lab is focused on a variety of topics, notably including those dealing with speech and language. One such application, Builder Bot, is a part of the “Metaverse”, and would allow users to change their virtual surroundings with voice commands. In terms of language, the lab is developing a “Universal Speech Translator”, which would be capable of instantaneous speech-to-speech translation between over 200 languages. Additionally, its “No Language Left Behind” program aims to provide text translation for over 200 languages as well, potentially allowing those who use low-resource languages to both access and share content in their native tongue. Notably, many of the projects under development by the lab are being shared as open-source, with the hopes of allowing others to build upon and improve the applications.

This AI lab, the first of its kind in The Netherlands, describes its goal as “accelerating the adoption of AI in Dutch society”. This lab provides consultation and training for organizations across all sectors, including private companies, governments, and NGOs. A quick look through its website illustrates their many and varied initiatives and projects, including “AI in Healthcare Program”, “Weight Gain Prediction for Pregnant Women”, and “AI for Good Hackathons”. Nonetheless, one initiative that stood out for us is “The Mothership”, an eighth-week long innovation program for teams and startups aiming to “build a community of individuals and organizations committed to helping vulnerable landscapes around the world”, and thus contribute to reach the UN Sustainable Development Goals. The initiative is co-founded by WorldStartup and Space4Good. It will be interesting to see how this lab evolves and inspires similar projects around the world.

Baidu, Inc. is one of the largest AI operations in the world. Co-located in Silicon Valley, Seattle, and Beijing, Baidu Research brings together top talents from around the world and large amounts of R&D spending to focus on future-forward research on AI, which has resulted in Baidu becoming one of the leading AI innovators in the world. From 2018, Baidu is part of an AI ethics group (Partnership on AI (PAI)). Its expertise include among many other topics, including the world’s first large-scale AI model dedicated to aerospace, as well as Digital Human, empowered with disruptive AIGC ability to reach new levels of artistic creation and interactive conversation. Additionally, Kaiwu industrial internet platform, based on Baidu AI Cloud, makes digital transformation possible across industries, and even farming communities. Another product is Apollo RT6, a fully autonomous vehicle that will provide driverless robotaxi service at half the price of traditional taxis. Baidu also made very important contributions to the control and prevention of COVID-19, including predicting passenger’s body temperature, vaccines design, etc. Last month, Baidu announced its first superconducting quantum computer that fully integrates hardware, software, and applications and the world’s first all-platform quantum hardware-software integration solution that provides access to quantum chips via mobile app, PC, and cloud. Baidu aims to build an open and sustainable quantum ecosystem where “Everyone Can Quantum”.

When the DNA sequence of the human genome was released in 2001, it was hard to imagine that a decades-old problem would be solved 20 years later. This breakthrough could be considered similar in magnitude to the discovery of the Higgs Boson, the verification of gravitational waves, or black holes. This effort, known as AlphaFold, is the most complex AI system from DeepMind, a research lab founded in 2010 and acquired by Google in 2014. But what does it mean to know the 3D structure of proteins? In biology, as in many areas, the structure of something tells us not only about its function, but about its mechanism of action. Inside the cell, proteins are synthesized from the reading and translation of genes. Once synthesized, they adopt a 3D folding that will determine their function. The leap from knowing the DNA sequence of genes and knowing the structure of proteins represented an astronomical advance. The effort to know the structure of a single protein could take years. Generations of doctoral researchers spent 4-5 years predicting such 3D conformations using methods such as crystallography and X-ray diffraction. Today, AlphaFold has reduced that time to seconds or minutes. This year, DeepMind announced the 3D prediction of more than 200 million proteins, representing virtually the entire known protein universe and probably the most important contribution AI has ever made to scientific knowledge, says DeepMind co-founder Demis Hassabis. The database of these predictions and the DeepMind computer code is freely accessible. With this advance, a range of possibilities opens up in the design of disease treatments, better food safety strategies or epidemic control, among many others. This effort also calls for coordinated work from different fronts regarding human rights, ethics and security. DeepMind’s work undoubtedly represents a moment to celebrate.