From now on, companies across the EU must be clear about their use of AI.

[eucommission@ec.social-network.europa.eu]

From now on, companies across the EU must be clear about their use of AI.

Providers and deployers of AI must now

️ Label AI-generated content as such.
️ Flag use of deepfakes
️ Inform users when they interact with AI or when it is used to analyse their behaviour.

️We have published guidance to help companies comply.

Transparency helps build trust

https://link.europa.eu/gTvhRX

[xsterbeprozessx@mastodon.social]

@EUCommission man tut so als wenn man damit den Menschen hilft, aber das ist eine massive Fördeurng der KI denn so kann jetzt automatisch KI Generierter Content aus den Training ausgeschlossen werden, die Folge bessere Deepfakes, bessere Modelle die Menschen nachahmen können, man hat das größte Problem des KI Trainings gelöst damit, wie kann ich KI Generierte Daten aus den Training ausschliessen. War es früher mir viel Aufwand verbunden ist es jetzt nur noch eine IF verzweigung

[fenrasulfr@mastodon.social]

@EUCommission Although this will not stop bad actors, it will help.

[hipsterelectron@circumstances.run]

@EUCommission so regulating false statements is out but they just have to be flagged as false statements that's cool very futuristic and forward-thinking

[robinadams@mathstodon.xyz]

@EUCommission A big step in the right direction.

[xgebi@hachyderm.io]

@EUCommission may I ask how it is enforced? Please, tell us that Ireland won’t be doing all the work like with privacy

[asardinianabroad@mastodon.social]

@EUCommission how could you pretend that deep-fake creators will flag their creation as a deep fake? And how will you manage the controls?

[lordtaku@mastodon.social]

@EUCommission ok, now we need to ban it's use and stop the construction of data-centers

[max@poliverso.org]

@ASardinianAbroad @EUCommission

There are a couple of replies in the comments above yours:

"A big step in the right direction"

and

"Although this will not stop bad actors, it will help".

[the_universality@mastodon.novotnykrystof.com]

@EUCommission Is EC & EP prepared to enforce this (and is there any will to enforce this)?

I'm asking because regarding the game industry, games that comply with Steam AI disclosure policy upon disclosing AI use report lower customer interest.
I believe you can see here in the replies how people hate AI so it isn't unrealistic for companies to not disclose it.

So, are we prepared or even capable of enforcing it?

[kasilas@mastodon.ie]

@xgebi @EUCommission this is almost pointless for exactly that reason - it will be left to national enforcement of loose law i.e. mainly looking away.

(That said, Ireland's enforcement of privacy rights is not much to brag about. When companies don't respect privacy law, as many based here don't, we almost never take action and the rare triggered action are slow and mere speeding fines when we do.)

[guepeethoux@piaille.fr]

@EUCommission We can't trust the tool itself, so ban the tool !

[pascaldragon@metalhead.club]

@XSterbeProzessX @EUCommission nicht wirklich, da der es nur für kommerzielle Nutzung von KI gilt. Private Anwender können weiterhin Inhalte ohne Kennzeichnung raushauen.

[xsterbeprozessx@mastodon.social]

@pascaldragon
hoffen wir mal das es so ist, wobei kommerzielle Nutzung ist schon die Generierung auf einer der vielen Plattformen eine solche?

ich würde es ja auch begrüßen wenn es wirklich nur für Menschen nutzbar wäre aber da es auch Maschinen nutzen können sehe ich da schon ehr auch die Gefahren.

@EUCommission

[ozzelot@mstdn.social]

@max
And both of those beg a simple question: "how?"
@EUCommission @ASardinianAbroad

[max@poliverso.org]

@ozzelot

I guess it'll follow the usual rule: you find something that breaks the law, you report it to the authority. If it's important they'll make an investigation and if the company is found guilty it'll be fined.

[prahou@merveilles.town]

@ozzelot @max @EUCommission @ASardinianAbroad

[kdelta@ieji.de]

@EUCommission So does that mean you’ll start doing it when you keep using terrible AI rendered images on here?

[vekkq@social.vivaldi.net]

@EUCommission AIs are considered complex systems, but the underlying algorithms are simple. AIs will simply be changed a bit to dodge the definition in the AI Act.

[gins@mastodon.gamedev.place]

@EUCommission So you're allowing companies to analyse user's behavior with "AI" as long as they mention it? Just like the cookie banner this will just abuse the majority of not tech savvy users who just blindly accept everything without being aware of what's really being done to them and give companies another get out of jail card, while doing nothing for users. How does a EU citizen exactly benefit from being analysed by "AI"?