There are endless guidelines, principles and directions for how to produce more ethical, inclusive, fair, robust, trustworthy, sustainable, secure or privacy-enhanced AI. We cut through the noise and know what’s fluff and what’s manageable and compliant. We rely on tried and tested approaches to AI auditing and ethical product development, and are not easily distracted by the newest trend in AI assessments
-
Founded in 2026 Gaard & Sekara (short for Skadegaard and Sekara), is a consultancy focused on assessing and auditing AI.
We take a harm-reductionist stance on AI products. In our experience, all AI, even use-cases believed to be low-risk, no-harm or “AI for Good”, require thorough risk-assessing. We understand algorithmic and machine-learning products as potential accelerators of existing inequities, and we often find biases and unfairnesses to reproduce in these systems even when not intended or built into the models, often through proxies.
-
Our team shares backgrounds that merge academic excellence with hands-on experience working on the security, safety, robustness, predictability, and auditing of AI and Machine-learning. In addition to holding PhD’s in a wide range of topics, from auditing AI, to modeling AI, to AI security and Representation, our team have all worked hands-on with developing, auditing, securing, improving, and assessing AI and Machine-Learning products. We have covered private and public sector use-cases across the globe – and we are used to working with a very wide variety of people; developers, designers, programmers, auditors, lawyers, you name it. We are all critical thinkers with a strong drive for making better, fairer, and more robust AI. We have all crossed back and forth from academia to corporate or public sector, and we all have international perspective and experience.
-
At Gaard & Sekara we advise those who buy, develop, and sell AI. We also assess AI products from inception through implementation. From Fortune 50’s to small businesses, from the UN to local municipal projects, our consultants are tried and tested across contexts. What matters is that you truly want to commit to reducing harm in your AI products and & data pipelines.
Our Team
Tess Skadegård Thorsen, PhD, Co-Founder
Award winning researcher and reluctant tech ethicist with a low tolerance for shitty AI.
Tess (she/her) is a reluctant AI ethicist and unlikely technologist. She is wired to critique and find risk. A diplomatic leader, Tess doesn’t compromise on truth or justice. Perhaps that is why her advice is sought out as an international consultant for companies and organizations, and as a board member for multiple organizations, including the Danish Pioneer Centre for AI. A former award-winning researcher (with a PhD in Politics & Society focused on representation), Tess regularly lectures and speaks across the globe for some of the Worlds most renowned universities and biggest companies. She was headhunted to run the first-of-its-kind Product Ethics Team at Comcast NBCUniversal in 2022. As a director at a Fortune 50 Tech and Media Conglomerate (at 32), Tess worked with a wide range of people on an array of questions – from product development, policy, legal, to investments, innovation, and incubation. She also attended the NYU Institute for Public Interest Technology to sharpen her coding skills – enabling her to work with the most advanced ML and algorithmic teams.
Regardless of what you are developing or where you are in your AI journey, Tess always knows the right questions to ask to uncover risks, inequities, and imbalance. Whether encountering a team of highly technical developers or a suite of top-level executives, Tess knows how and what to probe to rethink processes and make more robust and critical decisions (and products) along the way. She has the business-acumen to identify potentials and risks and the technical insight required to troubleshoot with even the most advanced teams.
As former VP of research at Comcast (Tess’ former boss), Patrick Cook explains “She is very generous in sharing her knowledge, never talks down, and always approaches collaborations with colleagues with a view to educate and, above all, inspire.”
Vedran Sekara, PhD, Co-Founder
The least professory AI-auditing professor you’ve ever met.
Vedran (he/him). A physicist with a PhD in Applied Mathematics, focused on computer science, Vedran did Machine Learning and Algorithms before it was hot and ran studies with some of the biggest human datasets science had seen. Vedran has had stints in both the public sector (leading machine learning research at UNICEF Innovation and advising the UNICEF Innovation Venture Fund) and the private sector (building machine learning pipelines at SONY Mobile). Today he splits his time as Associate Professor at the IT University when he isn’t crunching algorithms for Gaard & Sekara. He is an expert on social behavior and algorithmic robustness and his research has been covered in The Ecomomist, Forbes, Scientific American and Die Zeit, and been featured on the cover of Proceedings of the Natural Academy of Sciences.
Vedran can solve most problems you throw his way. He says he is a lazy thinker, and that’s what makes him think up fast and smart solutions that require less work (and less computing) – but to those who work with him he is always hailed as tireless and above all stubborn. He will not let a puzzle go unsolved: whether you are looking to challenge an ongoing issue through analytics of complex datasets or you want to ensure the robustness of an algorithm.
Vedran is impatient with problems, but exceptionally pedagogical with people. He loves partnering with a diverse team, drawing on everyone’s expertise. He has upskilled and collaborated with folks ranging from: MIT professors, branding teams, Google Brain scientists, government officials, to humanitarian aid workers and lawyers. He is measured and listens before making any calls, but as his colleagues from UNICEF always said: he will let you know when something is shit. His bullshit detector is unmatched. He even teaches a lecture on detecting bullshit.