The use of facial popularity for surveillance, or algorithms that manipulate human behaviour, will be prohibited in the proposal for an EU Artificial Intelligence Regulation.
Large-scale proposals, which were leaked prior to their official publication, also promised strict new regulations for what are high-risk.
This includes algorithms used through police and recruitment.
Experts said the regulations were flawed.
The use of AI in the military is exempt, as are systems used through government for public safety.
The list of prohibited AI systems includes:
European political analyst Daniel Leufer tweeted that definitions are very open to interpretation.
“How do you pay someone?he wrote.
For AI considered high risk, Member States apply much more oversight, adding a desire to appoint evaluation bodies to test, certify and inspect such systems.
And all corporations that expand prohibit or do not provide correct information about them threaten fines of up to 4% of their overall turnover, similar to fines for GDPR violations.
Examples of high-risk AI include:
Leufer added that the proposals “will be extended to include all public sector artificial intelligence systems, regardless of the assigned risk point. “
“This is because other people still don’t have the option to interact with an AI formula in the public sector. “
In addition to requiring new AI formulas to have human monitoring, the EC also proposes that high-risk AI formulas have a so-called closing switch, which can simply be a closing button or other procedure for closing the formula if necessary.
“Artificial intelligence providers will be incredibly targeted at these proposals, as this will require a basic replacement in the design of artificial intelligence,” said Herbert Swaniker, Clifford Chance’s lawyer.
Meanwhile, Michael Veale, professor of virtual rights and regulation at University College London, has highlighted a clause that will require organizations to reveal when deepfaux occurs, a debatable use of artificial intelligence to create fake humans or manipulate photographs and videos of genuine people.
He also told the BBC that the law was basically “aimed at distributors and experts who promote – absurd – the generation of AI in schools, hospitals, police and employers. “
But he added that generation corporations that used AI “to manipulate users” would possibly also have to replace their practices.
With this legislation, the EC has had to walk the tightrope between making sure the AI is used for what it calls “a toolArray . . . to the ultimate goal of human well-being” and to ensure that it does not save EU countries, the United States and China, from technological innovations.
He said AI is already reporting many facets of our lives.
The European Centre for the Law on Non-Profit Organisations, which had contributed to the European Commission’s White Paper on AI, told the BBC that there is “a lot of vagueness and loopholes” in the proposed legislation.
“The EU technique for the definition of a maximum threat in opposition to low threat is, at best, productive, careless and, at worst, dangerous, as it lacks the context and nuances necessary for today’s complex AI ecosystem.
First, the Commission deserves the dangers of artificial intelligence systems in a rights-based framework, such as dangers to human rights, the rule of law and democracy.
“Second, the commission deserves to reject a simplified low-threat design and a staggered technique for AI threat levels. “
Details can be replaced before regulations are officially disclosed next week, and are unlikely to be enacted for several years.
Mexico’s collapsed viaduct had concerns beyond safety
Questions about the history of protecting a Metro line in Mexico City increase after a fatal collapse over an overpass.
Biden unveils plan to vaccinate American teens
Bill and Melinda Gates” agreed on a percentage of ownership”
Anger rises as Covid rages in Modi’s constituency
Why 2021 loaths like Dove Cameron’s year
Americans doubt Covid Video Vaccine Americans doubt Covid vaccine
Within a microchip race to meet demand. Within a microchip race to meet demand
The harmful Indian charlatan Covid
Is the Five Eyes alliance in trouble?
The old board that hooked the ancients
From food giants to packaging concerns
How emerging marine grades threaten my home How the sea point threatens my home
Letter questionnaire
You mean those songs?
Warming
What happens to your painting in excessive heat?
Bill and Melinda Gates” agreed to a succession division ”1
Belgian farmer moves French border2
What happens in this image of two American presidents? 3
France threatens to cut off power in Jersey
Madrid’s leader re-elected in a bitter Spanish vote5
The cleaner’s resignation causes an avalanche of support6
Mexico’s collapsed viaduct struggled beyond protection7
Five after machete attack in Brazilian nursery8
Bill and Melinda Gates divorce after 27 years9
Duchess of Sussex to the children’s book10
© 2021 BBC. The BBC is guilty of the content of external sites. Learn about our technique for external links.