And Jeremie asked a very interesting question about why people take the risk of publishing fake scientific papers. He might be interested to know that there's another side to this problem: why do people take the risk of publishing such articles? Last January, Science published a fascinating article on this subject:
Regarding Google's fine, the title from your source may be a bit misleading. The Autorité de la Concurrence gives more details about its decision on its Web site (https://www.autoritedelaconcurrence.fr/en/article/related-rights-autorite-fines-google-eu250-million). You will read that Google is not fined only because it used content from press agencies to train Bard, but because this proved that Google breached its commitment to cooperate with a monitoring agency and to comply with commitments following a previous decision from the Autorité regarding how Google breached a legislative framework "aiming to create the necessary conditions for balanced negotiations between press agencies, publishers and digital platforms" some years ago.
In short: this case is really specific, as it concerns a pre-existing legislative framework and Google's commitments to respect it, and it is not really possible to tell how much Google has been fined for training Bard with the content of press agencies without their consent. In the US, the legal challenge should therefore be somewhat different.
Chapter 4 of the decision is worth reading, because it shows what kind of proof has been taken into account. The Autorité argues that when asked, Bard provides its user with the references of articles which it claims to have read (it's as simple as that!). Also, it asserts that the "Google-Extended" program, which should allow to opt-out, did not work for two major news web sites.
And Jeremie asked a very interesting question about why people take the risk of publishing fake scientific papers. He might be interested to know that there's another side to this problem: why do people take the risk of publishing such articles? Last January, Science published a fascinating article on this subject:
https://www.science.org/content/article/paper-mills-bribing-editors-scholarly-journals-science-investigation-finds
Great one, cyber-podcasters.
Regarding Google's fine, the title from your source may be a bit misleading. The Autorité de la Concurrence gives more details about its decision on its Web site (https://www.autoritedelaconcurrence.fr/en/article/related-rights-autorite-fines-google-eu250-million). You will read that Google is not fined only because it used content from press agencies to train Bard, but because this proved that Google breached its commitment to cooperate with a monitoring agency and to comply with commitments following a previous decision from the Autorité regarding how Google breached a legislative framework "aiming to create the necessary conditions for balanced negotiations between press agencies, publishers and digital platforms" some years ago.
In short: this case is really specific, as it concerns a pre-existing legislative framework and Google's commitments to respect it, and it is not really possible to tell how much Google has been fined for training Bard with the content of press agencies without their consent. In the US, the legal challenge should therefore be somewhat different.
The full text of the decision is available in French here: https://www.autoritedelaconcurrence.fr/fr/decision/relative-au-respect-des-engagements-figurant-dans-la-decision-de-lautorite-de-la-0
Chapter 4 of the decision is worth reading, because it shows what kind of proof has been taken into account. The Autorité argues that when asked, Bard provides its user with the references of articles which it claims to have read (it's as simple as that!). Also, it asserts that the "Google-Extended" program, which should allow to opt-out, did not work for two major news web sites.