In recent days there have been numerous movements around the launch of a series of initiatives to regulate artificial intelligence. Sam Altman, CEO of OpenAI, testified before the US Congress to give his assessment on the matter, and meanwhile there has been preliminary approval of the IA Act, the regulation of the European Union. That poses potential problems for ChatGPT users in the old continent.
The one who warns is not a traitor. According to the Financial Times, Sam Altman has warned that Brussels’ efforts to regulate the field of AI could force OpenAI to withdraw its services from European territory. That would mean that users from EU member countries would not be able to access ChatGPT, DALL-E 2 or Whisper.
If they can’t deliver, they’ll shut down in Europe. The comments came during Altman’s visit to the United Kingdom, one of the stages of a “tour” that has also made him visit spain. In London, the CEO of OpenAI explained that he had “many concerns” about the AI Act. This regulation aims to include the latest LLMs, such as GPT-4, in that legislation. For Altman “the details matter. We will try to comply [con la regulación]but if we can’t we will stop operating [en la UE]”.
An increasingly restrictive regulation. The AI Act was initially designed to monitor the most risky use cases of AI, such as its use in medical equipment or in financial loan systems or personnel hiring. However, the advance of platforms such as ChatGPT or Midjourney has caused the EU to expand the text and regulations. Among the measures, it would force the creators of these models (such as GPT-4) to identify and avoid the risks that this technology can cause in various scenarios.
Who is responsible for what. In essence, what this AI Act requests is that OpenAI (for ChatGPT) or Google (for Bard) take responsibility for what users or other companies do with these tools even when neither one nor the other has control over those particular applications of this technology. These reference companies would also have to publish summaries of the copyrighted content they have used in their AI models, thus opening up the possibility for copyright holders to claim compensation.
Europe against Big Tech. Peter Schwartz, a manager at Salesforce, explained that the lack of European benchmarks in the development of AI once again places the continent in a disadvantaged position as we already had in other fields such as the development of microchips. “Basically, it will be European regulators that regulate US companies, as has been the case throughout the information technology era,” he told the FT.
Sundar Pichai campaigns. Like Altman, the CEO of Google has taken a journey in which he has met with various political representatives and regulators of the European Union. Among them with Thierry Breton, responsible for the AI Act. Pichai told him about an “AI Compact” that could serve as initial regulations for companies that could then better adapt to that future formal regulation. Google officials recently published a statement on his official blog talking about his willingness to contribute to that regulation, as did those responsible for OpenAI in a similar message signed by Altman these days. However, the EU must be careful: if it is too strict, it could leave its citizens without access to these tools.
Image | Global Village | jernej furman
In Xataka | This is how I automated my entire house with artificial intelligence and an open source system