In brief European lawmakers recently added a clause to the AI Act – legislation proposed to regulate machine learning systems – requiring AI developers to disclose copyrighted data used to train their models.
The inspection of copyrighted data used to train generative AI models was reportedly added to the bill just two weeks ago.
“Against conservative wishes for more surveillance and leftist fantasies of over-regulation, parliament found a solid compromise that would regulate AI proportionately, protect citizens’ rights, as well as foster innovation and boost the economy,” said Svenja Hahn, a European Parliament deputy.
The AI Act is the first major government effort to regulate the technology with new laws. It was first drafted over nearly two years ago and is still being debated and finalized.
Efforts to start regulating AI in the US are also underway. Senator Michael Bennet (D-CO) introduced legislation this week to create a specialized cabinet-level task force to work on developing policies to ensure AI models uphold US laws protecting civil liberties.
Republican Party creates bizarre anti-Biden advert using AI
Hours after US president Joe Biden announced he will be seeking a second term in the White House, the Republican National Committee released an eerie anti-Biden advert completely generated using AI.
The video is made up of strange images that look like blurry photographs. If you stare hard enough, you can see some unsettling features suggesting they have been faked. Faint white text in the top left corner confirms that the advert was “built entirely with AI imagery”.
We were warned about this. When machine learning models started getting better at generating realistic images, experts warned that the technology would be used to create fake content in political campaigns. This is the first time that a major political party in the US has turned to AI to create an official advert against a rival politician.
The Democratic National Committee hit back against the video in a statement to The Washington Post, saying the RNC “had to make up images because, quite simply, they can’t argue with president Biden’s results.”
Self-driving biz Cruise running driverless cars in San Francisco 24/7
Cruise’s autonomous driverless cars are currently operating in San Francisco 24/7, the startup’s CEO, Kyle Vogt, confirmed this week.
“Operating robotaxis in SF has become a litmus test for business viability. If it can work here, there’s little doubt it can work just about everywhere,” he wrote in a tweet.
Vogt said Cruise will start deploying its driverless robo-taxi fleet in other US cities, and that the company’s machine learning systems powering its autonomous driving capabilities trained in San Francisco have “proven themselves” elsewhere too.
The areas that its cars can operate are more restricted during the day – they avoid busier neighborhoods closer to downtown, for example – but can drive around most of the city at night.
Officials from San Francisco’s transport agencies have warned against expanding self-driving cars too quickly – especially since they can block traffic, disrupt emergency workers, and have even collided with public transport vehicles.
Anthropic strikes deal with Scale to commercialize Claude language model
Anthropic has partnered up with data-labelling startup Scale to deploy and manage commercial generative AI applications.
Scale’s customers can build products and services powered by Claude, Anthropic’s AI chatbot assistant. They will be able to import their own datasets and spin up the software in a private AWS environment.
“Partnering with Scale allows us to bring our useful model, Claude, to more customers in a thoughtful, scalable way,” Anthropic’s CEO, Dario Amodei, explained in a statement.
“By combining Scale’s AI engineering capabilities with our values-based model development approach, customers have more assistance as they build and deploy generative AI applications. I’m excited about this partnership and the work we’ll do together to positively shape the future of AI.” ®