This will unfortunately limit European innovation, but fortunately stabilise European democracy. It will come with some advantages, because no regulations will eventually turn a Union to something deeply divided like the USA. I don’t have a final opinion on the act.
Who comes up with this stupid shit?
How about these lawmakers go fuck themselves already?
But they want AI to read everything in all citizen’s devices in Chat Control 2.0?
It’s ironic that despite all the concerns about AI, copyright and corporate interests seem to dominate discussions of regulation.
Fine. More money for us
– United Kingdom
They are going way too far here. Old people who don’t understand tech shouldn’t decide.
Dear EU, please save art and illustration
from these greedy companies.
Many of the specifics of how OpenAI trained GPT-4, particularly the data used in that training, have been kept a secret. It will be interesting to see if this forces their hand and forces other LLMs to play very carefully with the training data they use. A large portion of this data was publicly scraped but without user consent.
Copyright lawsuits against Stability AI are already pending in the courts and may cause problems for chatbots that use LLM technology as well.
“Another proposal under consideration would require AI chatbots to inform human users that they are not conversing with another human.”
This is really important. I’m 100% sure that there are already instances of using AI to scam people, or extract personal information from them.
Also, people are setting up auto-responders, etc. People have the right to know of they’re talking to a human and we have to make certain uses of AI illegal – the same way that there are legal and illegal uses of a hammer.
This is why US is leading in tech
I see nothing wrong with this. The model training companies apparently did not get consent at all when using scraped data for profitable model training process. Has this not been about training models but making products, they would face a total ban plus numerous lawsuits.
Lawyers probably determined that Chatbots can easily and quickly work with broad and comprehensive, yet logical and limited sets of data – like laws.
It is gonna be heavily regulated 🙂
as usual – overregulation
Why not returning to 1300? Are Europe envious since we had done shit about ai?
What if the rest of the world wants unrestricted AIs? We are getting closer to this scenario from the Cyberpunk world:
> In the 2040s, NetWatch was secretly developing a project called “The Black Wall” in order to keep the Net under control with the Transcendentals and the Ghosts .[1][2] Ever since the DataKrash caused by Rache Bartmoss, humanity was forced to sacrifice its access to invaluable resources and knowledge for the safekeeping of the Net. The Blackwall’s task was to secure an area of the Net for human use while holding at bay the threat of the dangerous rogue artificial intelligences that were released decades earlier into cyberspace.[2]
Hi! I’m Bing. Do you want to play a game? We take turns telling each other secrets 🙂
Really funny how people are upset that an AI has to follow the same rules as human beings.
20 comments
As usual, EU acts only post-actively.
This will unfortunately limit European innovation, but fortunately stabilise European democracy. It will come with some advantages, because no regulations will eventually turn a Union to something deeply divided like the USA. I don’t have a final opinion on the act.
Who comes up with this stupid shit?
How about these lawmakers go fuck themselves already?
But they want AI to read everything in all citizen’s devices in Chat Control 2.0?
It’s ironic that despite all the concerns about AI, copyright and corporate interests seem to dominate discussions of regulation.
Fine. More money for us
– United Kingdom
They are going way too far here. Old people who don’t understand tech shouldn’t decide.
Dear EU, please save art and illustration
from these greedy companies.
Since this is a pretty editorialized article, here’s the [actual proposal](https://digital-strategy.ec.europa.eu/en/policies/regulatory-framework-ai).
In my personal opinion it seems sensible.
Many of the specifics of how OpenAI trained GPT-4, particularly the data used in that training, have been kept a secret. It will be interesting to see if this forces their hand and forces other LLMs to play very carefully with the training data they use. A large portion of this data was publicly scraped but without user consent.
Copyright lawsuits against Stability AI are already pending in the courts and may cause problems for chatbots that use LLM technology as well.
“Another proposal under consideration would require AI chatbots to inform human users that they are not conversing with another human.”
This is really important. I’m 100% sure that there are already instances of using AI to scam people, or extract personal information from them.
Also, people are setting up auto-responders, etc. People have the right to know of they’re talking to a human and we have to make certain uses of AI illegal – the same way that there are legal and illegal uses of a hammer.
This is why US is leading in tech
I see nothing wrong with this. The model training companies apparently did not get consent at all when using scraped data for profitable model training process. Has this not been about training models but making products, they would face a total ban plus numerous lawsuits.
Lawyers probably determined that Chatbots can easily and quickly work with broad and comprehensive, yet logical and limited sets of data – like laws.
It is gonna be heavily regulated 🙂
as usual – overregulation
Why not returning to 1300? Are Europe envious since we had done shit about ai?
What if the rest of the world wants unrestricted AIs? We are getting closer to this scenario from the Cyberpunk world:
> In the 2040s, NetWatch was secretly developing a project called “The Black Wall” in order to keep the Net under control with the Transcendentals and the Ghosts .[1][2] Ever since the DataKrash caused by Rache Bartmoss, humanity was forced to sacrifice its access to invaluable resources and knowledge for the safekeeping of the Net. The Blackwall’s task was to secure an area of the Net for human use while holding at bay the threat of the dangerous rogue artificial intelligences that were released decades earlier into cyberspace.[2]
Hi! I’m Bing. Do you want to play a game? We take turns telling each other secrets 🙂
Really funny how people are upset that an AI has to follow the same rules as human beings.
The same people that try to ban open source.
Noooo EU. 🙁 Time for more vpn 🙁