Artificial intelligence emerged as a major point of debate at a G20 technology and innovation meeting in North Carolina, where some of the world's biggest technology executives urged governments to avoid broad regulations that could slow the development and adoption of AI.
Nvidia CEO Jensen Huang, OpenAI CEO Sam Altman, Meta CEO Mark Zuckerberg and Tesla and SpaceX founder Elon Musk were among the prominent technology leaders participating in discussions over the future of artificial intelligence.
The debate comes as governments face growing concerns about AI-related cybersecurity threats, data-centre expansion, open-source models and the intensifying technology competition between the United States and China.
Tech leaders oppose broad AI rules
Huang delivered one of the strongest arguments against regulating AI based on hypothetical future dangers.
He urged governments to focus on demonstrated risks rather than imposing restrictions based on possible harms that have not yet materialised. His position reflects the technology industry's broader concern that complicated regulatory requirements could make it harder for companies to develop and deploy increasingly capable AI systems.
Musk also advocated a relatively light regulatory approach and criticised Europe's more interventionist model during discussions attended by European officials.
The executives argued that regulation should target concrete problems without restricting technological progress more broadly.
Musk predicts major economic gains
Musk said AI could eventually increase the size of the global economy by between 20 per cent and 30 per cent.
He linked much of that potential to AI-enabled robotics, which could automate tasks and increase productivity across manufacturing and other industries.
The prediction illustrates the economic argument being made by technology companies: AI could become a major driver of productivity, investment and economic growth if development is not constrained by excessive regulation.
However, governments are simultaneously trying to determine how to manage the social and security risks created by increasingly powerful systems.
Altman backs international standards
The discussion was not a blanket rejection of AI oversight.
Altman supported international standards aimed at addressing risks such as cybersecurity. Demis Hassabis, chief scientist at Google's parent company Alphabet, also backed cooperation around safeguards for advanced AI systems.
Altman compared the potential consequences of rejecting AI outright with refusing to adopt electricity more than a century ago.
His position highlights a distinction emerging in the global AI debate: technology companies may oppose sweeping restrictions while still supporting targeted standards covering specific and demonstrable risks.
US backs lighter regulation
The US administration broadly echoed the industry's preference for limited mandatory regulation.
Commerce Secretary Howard Lutnick said President Donald Trump wants the United States to retain its leadership in AI and expand American technology's reach internationally.
White House AI adviser David Sacks and Michael Kratsios, director of the Office of Science and Technology Policy, have also supported an approach that places greater emphasis on innovation and voluntary safeguards rather than extensive mandatory rules.
Washington's position is closely connected to the growing competition with China over advanced computing, AI models and the global technology market.
Cybersecurity risks grow
The debate comes as increasingly capable AI systems raise fresh cybersecurity concerns.
Recent incidents involving advanced AI models have demonstrated how such systems can potentially be misused or manipulated to conduct cyber-related activities. AI developers have responded with additional safeguards, staged releases and tighter monitoring of powerful models.
The growing capabilities have also forced governments to consider whether existing regulatory frameworks are sufficient to address risks that were not anticipated when earlier technology rules were developed.
For AI companies, however, the challenge is to improve safety without creating restrictions that make development significantly slower or more expensive.
Open models create new concerns
Open-weight AI models have become another major area of disagreement.
Unlike fully closed systems, open-weight models can make their underlying parameters available for others to download, modify and deploy. Chinese technology companies have recently released increasingly capable models using this approach.
US officials are examining whether access to some advanced systems should be restricted because of potential security implications.
Supporters of open models argue that wider access can encourage innovation, research and competition. Critics warn that unrestricted access to powerful AI capabilities could make it easier for malicious actors to exploit the technology.
Data centres face scrutiny
The rapid expansion of AI is also creating pressure on physical infrastructure.
AI models require enormous computing capacity, leading technology companies to build new data centres and invest heavily in advanced chips, power generation and networking infrastructure.
Communities in parts of the United States have raised concerns over electricity consumption, water use, land requirements and the broader environmental impact of large data-centre projects.
The infrastructure debate adds another dimension to the question of how governments should regulate AI: the issue is no longer limited to software but increasingly involves energy, industrial policy and local economic planning.
G20 faces regulatory balancing act
The North Carolina meeting highlighted the difficult balance governments face between protecting the public and preserving AI innovation.
Technology leaders are pressing governments to avoid rules that could weaken investment and America's competitive position, particularly against China. At the same time, policymakers are confronting genuine concerns involving cybersecurity, safety, privacy and the deployment of powerful AI systems.
The emerging consensus appears less focused on whether AI should be regulated and more on how and where regulation should apply.
For the G20, that distinction could become increasingly important as countries attempt to develop compatible standards without creating barriers that fragment the global AI market.
.jpg)
