Artificial intelligence.. gets out of control
In the face of the ingenuity of artificial intelligence, warnings are multiplying even from some of those behind the development of GPT chat software, and as the European Union slowly moves toward its regulation, the technology's rapid rise seems to be overtaking all.
“We need global regulation ,” Sam Altman, head of OpenAI, told the US Senate last Tuesday, expressing his fear that this technology could cause massive damage to the world, by manipulating elections or causing disruptions in the labor market.
Like him, US tech giants demand to frame its use as they announce new AI services every week, so Christina Montgomery, IBM's vice president of trust and privacy, has called for regulating the sector without stifling innovation. And last April, Google President Sundar Pichai said: "In the long term, it will be necessary to develop a global regulation, but we are still at the beginning."
But the dissenters from the big tech companies, known as "GAVA", are the most ferocious, including Jeffrey Hinton, the 75-year-old co-founder of artificial intelligence , who quit Google in early May to raise awareness of the threat.
At the end of last March, more than a thousand personalities demanded a 6-month cessation of artificial intelligence research, including Elon Musk, who is also developing an artificial intelligence company, and thinker Yuval Noa Harari, who is convinced that artificial intelligence can destroy humanity .
"We don't want a world where 5 companies put humanity on the plane of artificial intelligence without thinking about the future we want," said activist Tristan Harris, Google's former chief ethics officer.
In Europe, the alarm is sounded above all by researchers, experts and regulators, and anti-AI rhetoric has multiplied in recent weeks.
Consultants from Vae Solis confirmed to the French newspaper Le Monde that "the production of false images and videos with artificial intelligence poses an unprecedented risk to the upcoming elections." "There is an urgent need to take back control of AI, which would cause unprecedented social unrest," said artificial intelligence researcher OG Bersini.
Meanwhile, the European Union is making progress on “artificial intelligence law,” and last week the European Parliament approved a framework that must be voted on in June. But it will take years before it can be put into practice, and the script will require generative AI systems such as ChatGPT to obtain prior permission and transparency of the algorithms and their data.
"Everyone agrees on the need for rules, even companies," says expert Ivana Bartoletti, director of privacy at consulting firm Wipro. She confirms that thinking is under way in the Council of Europe, the United States and the United Nations. "We need to avoid alarmist statements, which will only serve to frighten people rather than help them learn to use AI responsibly," she says.
But in Europe as in the United States, economic circles focus on the race for artificial intelligence for its economic potential as well as its geopolitical importance , as it is "an element of national security," says the CEO of Google.
In Europe, pro- AI corporations and politicians fear that the “artificial intelligence law” will do more harm than good, and Gilles Papineh writes in La Tribune: “We should take lessons from the European Personal Data Regulation, which instead of slowing down big digital companies, gave them an edge . ” Because for Gava companies, alone an army of lawyers is necessary. Papineh is the president of France's government-appointed National Digital Council.
"It's not a good idea to follow the example of Italy's data protection authority, which banned GBT Chat for 3 weeks because it collected data without authorization, and then re-licensed the US software after being compliant, but faces similar complaints elsewhere in the EU," he added.
“Europe cannot miss such an important technological breakthrough again,” said Cedric Au, former French Minister of the Digital Domain. The version of the AI law amounts to de facto banning the appearance of models in European languages for text analysis, translation, and machine speech. Without specifying the reasons, Google announced this . This week, Bard's competitor ChatGBT has been deployed in 180 countries, but not in the European Union.
Source : websites