Informação relevante
Don’t ask if artificial intelligence is good or fair, ask how it shifts power

It is not uncommon now for AI experts to ask whether an AI is ‘fair’ and ‘for good’. But ‘fair’ and ‘good’ are infinitely spacious words that any AI system can be squeezed into. The question to pose is a deeper one: how is AI shifting power?

Detecção de cyberbullying com restrições de equidade

O cyberbullying é um fenômeno adverso generalizado entre as interações sociais online na sociedade digital de hoje. Embora numerosos estudos computacionais se concentrem em aprimorar o desempenho da detecção de cyberbullying de algoritmos de aprendizado de máquina, os modelos propostos tendem a apresentar e reforçar vieses sociais não intencionais. Neste estudo, tentamos responder à pergunta de pesquisa "Podemos mitigar o viés não intencional dos modelos de detecção de cyberbullying, orientando o treinamento do modelo com restrições de equidade?".

Cyberbullying Detection with Fairness Constraints

Cyberbullying is a widespread adverse phenomenon among online social interactions in today's digital society. While numerous computational studies focus on enhancing the cyberbullying detection performance of machine learning algorithms, proposed models tend to carry and reinforce unintended social biases. In this study, we try to answer the research question of "Can we mitigate the unintended bias of cyberbullying detection models by guiding the model training with fairness constraints?".

Moldando o terreno da competição das IAs

Como as democracias devem competir efetivamente contra regimes autoritários no território das Inteligências Artificiais? Este relatório oferece uma "estratégia de terreno" para os Estados Unidos alavancarem a maleabilidade da inteligência artificial para compensar as vantagens estruturais dos governos autoritários na engenharia e implantação da IA.

Shaping the Terrain of AI Competition

How should democracies effectively compete against authoritarian regimes in the AI space? This report offers a “terrain strategy” for the United States to leverage the malleability of artificial intelligence to offset authoritarians' structural advantages in engineering and deploying AI.

Beyond a Human Rights-based approach to AI Governance

This paper discusses the establishment of a governance framework to secure the development and deployment of “good AI”, and describes the quest for a morally objective compass to steer it.

AI Fairness

Ruoss et al. published the first method to train AI systems with mathematically provable certificates of individual fairness. Full source code is available on Github.

The new IKEA Data Promise (Video: 10:37)

Companies discover data ethics as part of their corporate communications: IKEA promises to embed data ethics into all their processes.

Preprint: Alternative personal data governance models

The not-so-secret ingredient that underlies all successful Artificial Intelligence / Machine Learning (AI/ML) methods is training data. There would be no facial recognition, no targeted advertisements and no self-driving cars if it was not for large enough data sets with which those algorithms have been trained to perform their tasks. Given how central these data sets are, important ethics questions arise: How is data collection performed? And how do we govern its' use? This chapter – part of a forthcoming book – looks at why new data governance strategies are needed; investigates the relation of different data governance models to historic consent approaches; and compares different implementations of personal data exchange models.

How Big Tech Manipulates Academia to Avoid Regulation

A Silicon Valley lobby enrolled elite academia to avoid legal restrictions on artificial intelligence.

Challenging algorithmic profiling: The limits of data protection and anti-discrimination in responding to emergent discrimination

Interesting article using intersectional analysis do define and recommend mitigations to emergent discrimination in algorithmic systems.

Checklist for safe and responsible digital health research

My colleague Camille Nebeker at UCSD specializes in ethics with digital health – she's developed a checklist for “safe and responsible digital health research” - it's private but you can request a copy.

Recommendations on the Ethical Use of Artificial Intelligence by the Department of Defense

The United States Department of Defense (DoD) enduring challenge is to retain a technological and military advantage while upholding and promoting democratic values, working with our allies, and contributing to a stable, peaceful international community. DoD’s development and use of artificial intelligence (AI) reflects this challenge.
The Defense Innovation Board (DIB) recommends five AI ethics principles for adoption by DoD, which in shorthand are: responsible, equitable, traceable, reliable, and governable. These principles and a set of recommended actions in support of them are described in this document.

The (not so) Global Forum on AI for Humanity

Earlier this week, I traveled to Paris to attend the Global Forum on Artificial Intelligence for Humanity (GFIAH). The by-invitation event featured one day of workshops addressing issues such as AI and culture, followed by a two days of panels on developing trustworthy AI, data governance, the future of work, delegating decisions to machines, bias and AI, and future challenges. The event was a part of the French government's effort to take the lead on developing a new AI regulatory framework that it describes as a 'third way', distinct from the approach to AI in China and the United States.

EU guidelines on ethics in artificial intelligence: Context and implementation (PDF)

You may have seen the EU Guidelines on Ethics in AI. As I believe we are entering in a 'regulating AI debate', it's worth sharing it here.

The Tech Pledge

Take the Tech Pledge and join the movement: Make tech a force for good. Created by 150 people in technology during this years Techfestival.

Artificial Intelligence Governance and Ethics: Global Perspectives

AI is increasingly being embedded in our lives, supplementing our pervasive use of digital technologies. But this is being accompanied by disquiet over problematic and dangerous implementations of AI, or indeed, even AI itself deciding to do dangerous and problematic actions, especially in fields such as the military, medicine and criminal justice.

The Ethics of Personal Data: Human Rights, Agency Costs, Protection Rackets, and Privacy Wrongs

A gaping ethical dilemma at the very heart of the data protection ecosystem practically guarantees widespread privacy wrongs.

Perguntas feitas
Make AI optional?

My one major requirement for ethical AI: all services that utilize/incorporate AI modeling of me must allow me to optionally disable AI. I.e. opt out of the AI modeling of me. @Oguzhan Gencoglu mentioned however that this might not be possible.

Comment on the human rights impacts of algorithmic systems

The Steering Committee on Media and Information Society (CDMSI) invites comments from the public on one draft text that was prepared by one of its sub-ordinate bodies and is meant to be adopted by the Committee of Ministers in early 2020. The draft recommendation of the Committee of Ministers to member states on the human rights impacts of algorithmic systems was prepared by the Committee of Experts on Human Rights Dimensions of Automated Data Processing and Different Forms of Artificial Intelligence (MSI-AUT). The experts will meet again in September to review all comments received and to finalise the draft ahead of its review by the CDMSI in December. Comments should be provided through email.

Test-driving a couple of ideas

What’s your first reaction to the twin claims that super-human, bias-free AI (let alone AGI) is not a) morally optimal nor b) possible? (The two claims are separate but linked. The issue at hand is whether we should try to get rid of bias and hence: a) is it morally imperative to get rid of bias and b) is it even possible to get rid of it.)