EU’s AI Act might be bad news for small firms

The US is warning that the EU’s new AI Act could end up hurting smaller European businesses while giving an advantage to larger companies who can easily afford to meet the expensive compliance requirements. Bloomberg got its hands on some papers, revealing that the State Department is getting all worked up about the European Parliament’s take on the upcoming law.

They’re especially concerned about the regulations for large language models (LLMs), which are the backbone of most AI tools that create content. The analysis discovered that several of these regulations were “blurry or unclear.” Additionally, it expressed worries about the act’s emphasis on the hazards associated with creating AI models, rather than those linked to their actual utilization.

Will the new AI Act affect productivity?

Washington cautioned that the new rules might slash productivity, trigger job shifts, and put a damper on investment in research and development and business growth in the area, messing with the competitiveness of European businesses.

Insiders familiar with the situation told the newspaper that the US input has already been passed on to EU leaders.

On the other hand, European businesses have voiced comparable fears. In June, executives from several major companies in the bloc conveyed “significant worries” in a letter dispatched to the Parliament, Commission, and member states.

“The draft legislation would jeopardize Europe’s competitiveness and technological sovereignty without effectively tackling the challenges we are and will be facing,” said the letter.

Also Read: Here’s what EU’s 5 futuristic technologies will test in space

European firms not in favor of the new AI Act

The people who signed the letter include top dogs from major companies like Heineken, Carrefour, and Renault, as well as heads of tech firms such as Ubisoft, TomTom, and Mistral AI. They’re saying the AI Act will make companies bail out of the bloc, investors shift their money elsewhere, and slow down progress in Europe.

Their main worry comes from a recent tweak to the rules. On June 14, the European Parliament tacked on fresh demands for generative AI tools like ChatGPT, which, according to the signers, would trigger “unfairly high” costs for following the rules and risks for taking responsibility.

They’re cautioning that this will shove Europe even further back compared to the US in AI advancement. They add that this effect will stretch from the economy to culture, as big language models will be woven into everything from search engines to digital assistants.

“States with the most powerful large language models will have a decisive competitive advantage… Europe cannot afford to stay on the sidelines,” said the letter.

When will the new rules come into act?

Conversely, certain EU nations, such as Italy, have begun to regulate generative AI even prior to the act coming into force. In a July survey, European consumers expressed their belief that the technology ought to be subject to stringent regulations.

The survey revealed that AI-driven online search is one of the top three fascinating applications in all countries except Spain, where it comes in fourth. People are also stoked about AI being incorporated into healthcare diagnostics, roadside aid, and suggestions for flights and hotels. Yet, most of the European folks polled reckon that society isn’t quite prepared for AI. This feeling is particularly strong in France and Germany, voiced by 74% and 70% of the participants, respectively. 

The AI Act is set to kick in around late 2025 or early 2026. Facing pushback from the business world and concerns of potential extinction, it will need to strike a delicate balance between ensuring safety and fostering innovation.

Also Read: How research of attoseconds helped these physicists win Nobel Prize

Is there really a risk of extinction?

Some of Europe’s top tech whizzes teamed up with a worldwide squad of IT gurus to sound the alarm that AI might bring about the end of the world. The warning, put out by the non-profit Center for AI Safety, has gathered signatures from various business honchos, scholars, and public figures. 

This list features Sam Altman, the head honcho at OpenAI, Kevin Scott, the bigwig at Microsoft, and, um, the musician Grimes. Surprisingly, her ex-boyfriend Elon Musk was nowhere to be found, despite his extensive history of expressing worries about the sector.

A big chunk of the people who signed up are from Europe. Some of the notable names include Demis Hassabis, the CEO of Google DeepMind, who hails from London, Kersti Kaljulaid, the former president of Estonia, and Geoffrey Hinton, a British smarty-pants who won the Turing Award and recently left Google to yap about the perils of AI.

The announcement adds to the flurry of recent warnings about the life-and-death risks brought on by AI. Just in the past couple of months, big shots in the industry have asked for a halt in training supercharged AI systems due to concerns about the threats to mankind. Healthcare experts have requested a break in the development of artificial general intelligence. Musk has cautioned that AI might bring about “the end of civilization,” and Google boss Sundar Pichai has confessed that the perils “give him sleepless nights.“

Skeptics might point out that many of the people raising the alarm are also pushing back against any AI rules that could hurt their businesses.

Business leaders have some solutions in mind

On top of voicing their concerns, the business heads came up with some ideas. Their main proposal is to limit EU regulations to general principles using a risk-based method, managed by a specific body that can adjust to new advancements and risks. They emphasized that this procedure should be developed through discussions with the business sector.

The folks who signed also showed some love for parts of the AI Act. They gave a thumbs-up to compulsory safety checks for new systems, a standard label on AI-made content, and a responsibility to be careful when creating models. But these gestures didn’t exactly win the hearts of the lawmakers. Dragos Tudorache, who co-headed the making of the AI Act, quickly shot down the letter.

The good news for the writers is that they still have plenty of time to pen more letters. The AI Act isn’t predicted to take effect until 2026 at the earliest.