The European Union has tightened AI restrictions to require corporations to disclose their data sources, which will challenge tech industry norms.

Since Microsoft-funded OpenAI launched ChatGPT last year, generative AI apps that quickly generate text, graphics, and audio have attracted attention and investment. However, ethical training data sources have become a serious issue, such as using famous books and Hollywood movies without authorization.

Over the next two years, the EU will implement the newly enacted AI Act in phases, giving authorities time to enforce new requirements while businesses adjust. 

Key requirements under the new law include "detailed summaries" of information needed to train general-purpose AI models like ChatGPT, according to Reuters. The newly created AI Office will provide a uniform compliance template by early 2025 after stakeholder engagements.

Tech Firms Says AI Law Could Harm Industry

AI firms argue that sharing training data will compromise their competitive edge by revealing secret knowledge. Matthieu Riouf, the CEO of Photoroom, compared it to protecting a chef's secret recipe. These EU transparency regulations will affect AI startups and big giants like Google and Meta, which rely on AI for future developments.

Content producers have sued Google and OpenAI for using their work to train AI models during these advancements. Data usage, copyright, and compensation issues persist both in the United States and worldwide.

Notably, the AI Act classifies AI applications by social risk. High-risk AI applications like social score systems that aggregate data, predictive policing, and emotional recognition in workplaces and schools are specifically banned.

According to the EU AI law, risky AI systems include autonomous cars and medical equipment that may affect public health, safety, and basic human rights. AI algorithms in financial services and education are being examined for biases, per CNBC.

In light of the rapid developments in artificial intelligence, UN Secretary-General Antonio Guterres urged nations to prevent AI from worsening societal inequality. On Wednesday, he warned before the Council of the International Telecommunications Union in Geneva that technological advances exceed legal frameworks, Anadolu reported.

Read Also: Teachers, Victims of Explicit AI Deepfakes Made by Students

Guterres called for nations to "collaborate to prevent AI from perpetuating inequity." He noted that addressing the matter needs creative and swift solutions like technical assistance and investments in data infrastructure and talent development for underdeveloped nations.

The UN chief stressed the need for countries to collaborate and address digital access and governance gaps. He saw the September Summit of the Future as a crucial opportunity to promote worldwide collaboration.

Experts Raise Warn of an AI-Driven Environment Crisis 

In addition to copyright and security threats, experts raise concerns about AI's massive energy consumption and growing carbon footprint that poses an environmental crisis, Nikkei Asia reported.

According to University of Glasgow computing science professor Wim Vanderbauwhede, each query on platforms like OpenAI's ChatGPT, used by 200 million people a year, can require 50 to 90 times more power than a Google search.

Studies suggest that generating an AI image may use as much power as 522 smartphone charges. Making 1,000 such images emits as much carbon as driving 4.1 miles in a gasoline-powered automobile. Researchers predict that recent breakthroughs from OpenAI, Google, and Microsoft, such as text-to-video functionality, will increase energy consumption and emissions.

Data centers, which power digital infrastructures globally, drive AI's energy usage. The IEA predicts worldwide data centers will triple their power use by 2026, hitting 1,000 terawatt hours.

Despite these estimates, AI resources increasingly shape global economic strategy, prioritizing economic competitiveness above environmental concerns. In Asia, governments are expanding data centers to boost digital and technical leadership, emphasizing the need for energy-efficient AI solutions.

As the industry faces these issues, semiconductor advancements such as Nvidia's energy-efficient AI processors are critical. According to the report, the newest Blackwell GPU from Nvidia claims 25 times higher energy efficiency than previous versions, a major step toward sustainable AI.

Related Article: China Introduces Photonic Chip Capable of Processing Hundred Billion Pixels in Just 6 Nanoseconds