Summary
Highlights
Artificial intelligence has become omnipresent in politics, economy, information warfare, and daily life, impacting our routines at work, on the internet, and in our cars. France is hosting a global summit on AI to ensure that generative AI isn't solely dominated by American and Chinese players. AI is already indispensable, featuring on phones, computers, and in cars. It's used in magazines, with some even written entirely by AI, leading to concerns about job displacement. Virtual models and AI-created influencers are emerging, like Anardy, a tourist guide with over 11,000 Instagram followers. AI is also integrated into cars, responding to requests. One in three French people already experience AI's daily impact, for example, when shopping online, but distinguishing between virtual and real has become a challenge.
Michel Levy Provençal notes that while the AI wave was anticipated, its speed has been surprising. The emergence of Chinese AI 'Dipsic' quickly disrupted the landscape dominated by American AI, only to be followed swiftly by new models from OpenAI, showcasing the rapid pace of innovation and competition. Despite concerns about data scarcity for training AI, advanced AI models can now generate their own artificial data. The profitability of AI companies, like OpenAI's $5 billion loss last year, is compared to the early days of the internet where companies like Amazon were initially unprofitable before becoming successful.
France's global AI summit aims to establish the country and Europe as key players, but realistically, Europe lags behind. The US invests massively, with projects like 'Stargate' committing $500 billion over five years, dwarfing Europe's $56 million investment in open-source AI. This disparity highlights Europe's financial and product capacity challenges against American giants like Microsoft (investing $10 billion in OpenAI), Google, Meta, Amazon, and Apple. The Chinese government also prioritizes AI, with companies like Baidu, Alibaba, and Huawei closely linked to the state, as demonstrated by the 'Dipsic' AI.
Dipsic, a Chinese AI, has revolutionized the AI model by achieving comparable results to ChatGPT with 50 times fewer resources and without the latest generation chips. This open-source development, despite doubts about resource claims and accusations of data theft from OpenAI, has challenged American AI dominance and caused a $1 trillion loss in Wall Street. Dipsic's ability to run on less powerful hardware, making it more accessible, is a major advantage. However, it is subject to Chinese censorship, raising concerns about shaping global perceptions through AI.
While AI could become more energy-efficient, increased usage might offset these gains. AI's impact on employment is a major concern, potentially transforming work needs and capabilities. Tasks in administrative, industrial, and technical sectors, including software development and even creative fields like manga creation, are at risk. A report by the World Economic Forum predicts 300 million jobs could disappear by 2030, particularly in advanced economies. However, jobs requiring human interaction are more resilient, and AI is also expected to create more jobs than it eliminates, similar to past technological revolutions.
Europe's approach to AI is seen as overly focused on regulation, potentially stifling innovation, while other regions like Africa and Asia embrace AI for productivity gains. Europe faces a dilemma: how to regulate AI while fostering innovation. Emmanuel Macron advocates for a global governance framework for AI, but this conflicts with fierce competition, as seen with Donald Trump's repeal of Biden's AI regulation. The idea is to revive the Global Partnership on Artificial Intelligence (GPAI), launched by France and Canada in 2020. Despite Europe's limited investment compared to the US, significant investments like the UAE's plan for a giant data center in France (30-50 billion euros) and 35 data centers, leveraging France's decarbonized energy, could boost Europe's role in the AI value chain.
AI offers significant potential in healthcare, particularly in predictive medicine. It can detect diseases, like breast cancer, more effectively by analyzing vast amounts of data and comparing it to known cases. AI can also predict disease risk, enabling proactive monitoring. The development of AI like AlphaFold, which can discover how proteins function, points to a future where AI acts as a researcher in medicine. Making these advancements accessible to a wider population is crucial to prevent AI from exacerbating health inequalities, as preventive medicine currently remains a privilege for an elite.
The concept of AI transforming human knowledge into equations is not new, dating back to 1985. However, generative AI eliminates the need for human input, with machines capable of generating their own data. While current AI lacks complete autonomy, long-term memory, and planning capabilities, the goal is to achieve Artificial General Intelligence (AGI), which can rival human intelligence. Some experts predict AGI will arrive as early as 2026-2027, a key objective for companies like OpenAI.