Check out the submission from myself and Fanny Hidvégi from Access Now to the European Commission’s ‘White Paper on Artificial Intelligence – A European Approach.’ As well as contributing via the online consultation form, we have put together an additional document outlining our six key recommendations:
We must not promote the indiscriminate uptake of artificial intelligence. The uptake of any technology, particularly in the public sector, should not be a standalone goal and it is not of value in itself.
Implement a rights-based approach and mandatory human rights impact assessments. We oppose a binary high/low risk approach. All AI applications must respect human rights. We call for mandatory human rights impact assessments for all AI/ADM.
No safeguard or remedy could make indiscriminate biometric surveillance or predictive policing acceptable, justified or compatible with human rights. Ban applications that are incompatible with fundamental rights such as biometric technologies that enable mass surveillance.
Establish national centres of AI expertise to help existing regulators (recommendation in collaboration with AlgorithmWatch). Scope: monitor, assess, conduct research, report on, coordinate but no direct enforcement or complaint handling.
Establish public registers for AI/ADM systems (recommendation in collaboration with AlgorithmWatch)
Enforce high scientific standards
“The EU has an obligation to respect and promote human rights. The European Commission should put human rights at the center of artificial intelligence policies with legislation for mandatory human rights impact assessments and the creation of public registers.”, said Fanny Hidvegi, Europe Policy Manager at Access Now.
“Although strong regulation and safeguards can mitigate certain harms, we need to accept that certain uses of technology, such as biometric recognition systems that enable mass surveillance, are so incompatible with the protection of fundamental rights that these systems simply should not be used,” said Daniel Leufer, Mozilla Fellow at Access Now.
“On the other hand, Mozilla’s Daniel Leufer called the plans “disappointingly, if predictably, weak on facial recognition/ ai-powered biometrics. It acknowledges that the biometric processing involved is illegal and poses a threat to fundamental rights, yet only calls for ‘broad European debate.’””
Quartz reached out to me for a comment on the EU’s new data strategy for this article:
The full comment from myself and my colleague Estelle Massé at Access Now is below:
“Contrary to what Commissioner Breton said, the EU did not lose the so-called battle on personal data. In building the digital single market and enabling sustainable free flow of data, the EU became a leader in the protection of personal data globally. The EU approach on data, so far, has put people and human rights protections first, that same approach should be at the core of the new strategies. Instead of calling for more data and more AI, we need better and more responsible use of technology where it can really benefit people.” said Estelle Massé, Senior Policy Analyst at Access Now.
“There is a dangerously naive belief that ‘more data and more AI is always better.’ Collecting data and using AI to analyse it is not a neutral process; it can create new potentials for abuse, discrimination and surveillance. We often hear that China and the US are leading the so-called AI race because they have more data, more investments, and more AI uptake, but they also have the most infamous cases of AI abuses. Is this really a race we want the EU to enter? There are enormous unresolved issues around data security, not to mention the huge environmental impact of increased data storage which is at odds with any talk of a European Green Deal.” said Daniel Leufer, Mozilla Fellow at Access Now.
“The priority of the EU for the single market should be that all companies, American or not, respect human rights, including by complying with our privacy and data protection laws,” added Estelle Massé.
Beyond what’s quoted here, I also discussed the fact that we can’t really have the ‘broad European debate’ on this technology unless we hold off on installing the infrastructure that enables it. If we’ve got local governments signing contracts to install FRT-enabled cameras, it’s going to be hard to have a debate where the option of not using FRT can be properly considered.
The video from our panel on the ‘Liquid AI’ (a.k.a. transnational governance of AI) is now online. Check out the contributions from Sjoera Nas, Adam Harvey, Kristina Irion and me, and the excellent moderating of Bodó Balázs.
Check out my new blog post for Mozilla on hype, myths and inaccuracies about AI. For the next 9 months, I’ll be working with experts to develop resources to challenge these narratives and myths. Let me know what myths need busting, and who to work with!
On the occasion of the publication of the EU’s ‘Ethics Guidelines for Trustworthy AI’, I wrote a blog post with Fanny Hidvégi, Access Now’s European Policy Manager and member of the High-Level Expert Group on AI.