AI – although the potential setback due to the mass adoption of synthetic data – it is constantly progressing in its technique, while new models from all over the world are sprouting like mushrooms encouraged by fierce competition: but Is artificial intelligence also evolving in its regulation?
In contrast to a more globalized digital market with respect to international relations, It depends on the country. In the EU there are the strongest attempts at ethical, privacy and security regulation (such as the Digital Services Act). In the States, business is clearly the priority. In Asia, there are many nuances, realities and mixed approaches, which we often don't know well. So there are many grey areas where even the same people can come forward big tech.
Especially if it becomes an opportunity to enter the merit of relevant economic and political decisions.
Thus OpenAI has its say, with some regulatory proposals: but in what context does it propose them? What are these proposals and what are their purposes?
Content index
The Global Context of AI Regulation
Attempts to establish regulations come from everywhere, for different reasons. But (really or superficially) there is fundamental agreement on the need for common regulations on privacy and security.
We saw it at theAI Safety Summit, held in November 2023 at Bletchley Park, UK.
This event brought together participants from all continents between leaders, technology experts and company representatives. All with the aim of discuss the challenges related to advanced artificial intelligence and its regulation.
A historic event, which among its results includes the Bletchley Statement: un international agreement signed by 28 countries which highlights the need for cooperation between governments and businesses to ensure that AI is developed and used in a safe, reliable and transparent way.
OpenAI's proposals for regulating artificial intelligence
In this apparently constructive and proactive climate, OpenAI outlines proposals to regulate the development and deployment of advanced AI models. Line that has as its declared direction that of structuring a innovative regulatory framework at the same time ethical, sure, etc… but the crux of the matter seems seeking US government support to defend themselves.
One of the most relevant documents that gives you a better idea is theEconomic Blueprint, published by OpenAI in January 2025.
In this proposal, The company emphasizes the importance of more structured government support to maintain the United States' competitive advantage in artificial intelligence sector. According to OpenAI, in fact, public investments are fundamental for prevent the development of advanced technologies from being monopolized from a few companies or moving to other countries, like China, where the regulatory approach is less restrictive. So in essence he is “warning” the US that – if they do not support OpenAI – they will leave room for China.
DeepSeek Threat Drives OpenAI’s Proposals
China from where it came DeepSeek, which OpenAI publicly (but informally) accused of plagiarism, but which above all caused NVIDIA's stock market to collapse, demonstrating – essentially – that AI can be cheaper. But, rightly so, also too elusive and dangerous. It is not surprising, therefore, that among OpenAI's proposals there really is the need to regulate the AI critical infrastructure, as:
- chip production,
- access to data at scale,
- managing the energy resources needed to train the models.
The “Frontier AI” Plan
Then, OpenAI reiterated the importance of greater control over frontier AI models. Let's talk about the most advanced technologies that could have a significant impact on society if not managed well. The concrete proposal consists in creation of an independent body responsible for evaluating and certifying the respect for privacy and security of these models before they are released on the market.
Ma who is it made up of? this “independent body”? It is not yet known.
What is certain is that today OpenAI already has its own Security and Oversight Committee that on paper operates independently from the company, but which is composed of:
- Zico Kolter: professor and chair of the machine learning department at Carnegie Mellon University, currently chair of the committee.
- Adam D'Angelo: CEO of Quora and member of the board of directors of OpenAI.
- Paul Nakasone: General of the United States Army and director of the National Security Agency.
- Nicole Seligman: lawyer and former chairman of Sony Entertainment.
All present also in the Board of Directors of the company.
Proposals for the regulation of artificial intelligence in the field of security
In addition to asking for government support and protecting itself, OpenAI also proposes the creation of Traceability and certification mechanisms for AI-generated content. With the increasing spread of scams, deepfake and misinformation, is a crucial step to help users and platforms distinguish between authentic and artificial content. At the same time, OpenAI promotes partnerships with other companies and academic institutions:
- NIST: Collaboration to test and evaluate AI safety before public release.
- University of Virginia: integrating ChatGPT EDU into teaching.
- Faculty: supporting companies in the safe implementation of artificial intelligence.
- Anduril: Use of AI in drone countering systems.
Why does AI regulation matter so much?
Why pure technical evolution alone is reckless and why regulation opens up a broader vision of progress technological. But is this enough to answer? Obviously not.
Artificial intelligence fascinates, intrigues, excites, dazzles, assists, speeds up, optimizes; artificial intelligence scares, deceives, replaces, flattens, steals, lies. Sometimes it seems to us that we are dealing with something absolutely revolutionary, to find us disoriented and in one gray area, without any solutions other than utopian or dystopian extremes. Then, however, perhaps a small anecdote and a simple reflection are enough to clarify things better.
Has photography killed painting?
In 1839 The French painter Paul Delaroche saw it for the first time un daguerreotype: do you know what it is? We are talking about a silver metal plate where an image captured from reality is imprinted: in essence, the first form of commercial photography. Faced with a technology that was mind-blowing compared to what he was used to, he said: "From today painting is dead".
Tough statement, right? But understandable. Imagine a world where the only ways to represent something that others cannot reach is paint it, draw it, sculpt it or make her imagine through the writing. So he saw photography as an absolute threat for painting, but we know perfectly well how it went: currents of painting were born that revolutionized aesthetics, giving a contribution to culture that was far less didactic than “copies” of reality, often also devoted to forms of flattery and patronage.
And, what's more, thanks to photography you can see his works on the screen of a PC or smartphone. Which does not necessarily replace the enjoyment of the work: perhaps it can encourage it.
From this anecdote we can balance the most pessimistic and absolutist visions which concern the developments of artificial intelligence.
But that's still not enough for us.
Artificial Intelligence is our daughter
Of all the technologies ever conceived by humans in history, AI is the one we feel most like our daughter. So it is normal that we care even more that he “behaves well” and that he “grows up healthy”. Through the neural networks, we are recreating some mechanisms of the human brain… and this is just one example of the levels we have reached, and not even the most surprising one.
The step towards the Creating People with AI, or of “individuals” who affirm their own “identity”, is not so far from happening.
So, with the example of photography we can reflect on multiplicity of the impact of a technology, which is not always devastating and absolutizing; but we must keep in mind that we are actually dealing with the birth of “creatures” that we must nourish and educate carefully. The risks are as many as the opportunities.
So the regulation of artificial intelligence, becomes a compromise to be established as laws are established in all eras. And today, we are witnessing all the phases of this irregular, contradictory process, which leaves us breathless and full of questions.
