Source: Gelonghui

The big AI model is blinding and running wild, and regulation is rapidly stepping up.

Because AI training uses content from media publishers without permission, tech giants$Alphabet-A (GOOGL.US)$/$Alphabet-C (GOOG.US)$Recently, they have been severely punished in France.

Caught in a copyright dispute

On March 20, local time, the French competition authority fined Google 250 million euros (US$271 million).

The reason is that Google has violated EU intellectual property rules in its relationships with media publishers. This case mainly involved some of France's largest news agencies, including Agence France-Presse.

According to regulators, Google failed to reach an agreement with media publishers to post links to its content at will.

Additionally, Google's chatbot Bard (now known as Gemini) also uses media article content for AI training without permission, and there is no way to provide compensation or avoid infringement.

In fact, Google and media agencies have been involved in a long-standing dispute over “paying for news content.”

In 2022, the French Copyright Office issued a 500 million euro fine against Google and ordered it to negotiate licensing agreements with French publishers.

At the time, Google avoided fines by promising to negotiate in good faith with news providers on issues such as compensation for their content.

However, in a statement on Wednesday, regulators said Google had breached four of the seven promises in the settlement, including conducting honest negotiations with publishers and providing transparent information.

Although Google agreed to provide news agencies and publishers with a “transparent evaluation” of user rights rewards and ensure that the negotiations do not affect “other economic relationships” between Google and publishers.

However, regulators say Google has failed to deliver on these promises on several fronts.

First, it didn't share enough information with representatives to provide the information needed for the monitoring agreement in a timely manner; second, Google failed to provide full details on how to make money from news content, which also breached its 2022 promises.

No, the regulation “came to a head” with a fine of 250 million euros.

AI regulation

Nowadays, as AI sets off a global competition, copyright disputes have become a common problem facing big model training today.

According to reports, Google's Bard artificial intelligence service was launched in France in July 2023, but media companies were not told until September of last year that they used their content for AI training.

Regulators said Google promised not to dispute the facts. Despite reaching a settlement with regulators, Google responded that the amount of the fine was still too high.

“Our efforts to respond to and address concerns raised by interested parties have not been fully taken into account.”

However, Google said it would pay rather than raise an objection, and said “it's time to turn the page.”

In addition, Google also proposed a series of corrective measures aimed at resolving the discovered vulnerabilities. The regulators said they would pay attention to this.

In addition to Google being warned about AI training, OpenAI and Microsoft were also sued in court by the media earlier for AI training.

The New York Times was the first major media agency to sue ChatGPT and the AI platform over a copyright dispute.

The media believes that as millions of articles are used to train AI robots, these robots compete with news media and become sources of information.

It is worth noting that many countries are currently tightening supervision of AI.

In the middle of this month, the world's first “AI Act” was approved in the European Union. The bill is expected to take effect in May, and the legislation is expected to take effect in early 2025 and be implemented in 2026.

This unprecedented bill would ban certain AI applications that “threaten the rights of citizens,” including biometric classification systems based on sensitive features, and the untargeted capture of facial images from the Internet or CCTV footage to create facial recognition databases. AI that manipulates human behavior or exploits human weaknesses will also be banned.

Recently, Japan also began discussing the adoption of legal supervision against large-scale artificial intelligence (AI) developers, and plans to set penalties for situations where measures against false information have not been fully improved.

With the explosion of generative AI, many countries such as China, the United States, and the European Union are currently exploring the regulation of AI.

It can also be seen from this that in the surging AI wave, the EU's AI legislation is only the starting point.

editor/tolk