Is European regulation stifling innovation? Major tech companies have been singing this refrain in recent weeks. “Europe has become less competitive and innovative than other regions and now risks falling further behind in the era of artificial intelligence, due to inconsistent regulatory decisions”, accuse around thirty digital companies in this way, in an open letter published Thursday, September 19 and relayed by a petition, as well as advertising pages in major newspapers. At the head of this campaign: Meta, the parent company of Facebook and Instagram, whose founder, Mark Zuckerberg, had already co-signed at the end of August, with the boss of Spotify, Daniel Ek, a platform against regulation judged “too complex and incoherent.”
While not entirely unprecedented, this lobbying initiative is accompanied by a rarer decision: Meta has suspended the EU launch of its AI assistant, Meta AI, on its social networks Instagram and Facebook. In its letter, the group urges Brussels to stop “reject progress (…) and watch the rest of the world build technologies that Europeans won't have access to.”
While Apple has a less offensive communication than Meta and has not signed its open letter, Tim Cook's company has also suspended until further notice the launch in Europe of its AI assistant, Apple Intelligence, as well as three other features planned for the iPhone 16. To justify its decision, the group invoked the “regulatory uncertainties” generated according to him by the European regulation Digital Markets Act (DMA), some of which measures could force him to “compromising the integrity of [ses] products in a manner that endangers user privacy and data security.”
Meta is also protesting against a decision by the Irish CNIL which prohibits it, in the name of the general data protection regulation (GDPR), from training its AI models on content published on its social networks by its users, without asking their consent. “This is publicly shared data, not private conversations like messages on Messenger.”, argues Joëlle Pineau, vice-president in charge of AI research at Meta. It is important to train AI on European languages and culturess “so as not to have only Anglo-Saxon models”, adds the Canadian researcher. Meta finally argues that Brazil or the United Kingdom have given it the green light.
You have 66.81% of this article left to read. The rest is reserved for subscribers.
Source: Lemonde
