Summary of a Keynote from Cédric O at Les RIAMS.

Former French Secretary of State for Digital Affairs Cedric O recently gave a fascinating keynote speech at the Les RIAMS event, sharing his perspective on artificial intelligence, startup innovation, and the future of technology in France and Europe. As one of the co-founders of Mistral AI, a leading French AI startup, Cédric O has unique insights into the opportunities and challenges facing European technology. In his speech, he touched on a wide range of topics, from the motivations behind founding Mistral AI to the potential impacts of generative AI on cybersecurity. Cédric O’s candid thoughts provide a valuable window into the thinking of French political and business leaders as they grapple with both the risks and the promise of rapidly advancing AI capabilities.

Stepping Down from Politics to Return to Startups

Cédric O opened his keynote by reflecting on his surprising decision to leave his ministerial post in the French government and return to the private sector. As he put it, “People saw me climbing the ranks of government ministries but, I decided no, I’m done, I’ll take a break and do something else.”?This choice came as a shock to many who expected Cédric O to continue rising to higher levels of influence in politics.

However, as Cédric O explained, he never envisioned a lifelong political career for himself. He believed it was valuable for people from the private sector to bring their experience into government, and that they should then cycle back out into new roles. True to this vision, after two years away from politics, Cédric O has been heavily involved in launching Mistral AI , serving as one of the company’s co-founders.

Cédric O saw this return to startups as beneficial both professionally and personally, noting that since leaving his ministerial post he has had twins. As he quipped, “the logistics are easier when you’re not a minister.” While no longer playing an active executive role at Mistral AI, he continues to support and advise the company as it grows. In addition, Cédric O is exploring launching another AI startup focused on education technology.

Mistral’s Journey from European Dream to Microsoft Partnership

A significant portion of Cédric O’s speech focused on recounting the origins and evolution of Mistral AI. As he described, the company emerged out of a vision shared by three AI researchers, Guillaume Lample , Arthur Mensch , and Timothee Lacroix . United by a common background from prestigious French universities, the trio aimed to build an AI company capable of competing with American giants.

Initially, Mistral AI founders proudly proclaimed their intention to create a leading European AI company. However, Cédric O explained that partnerships with major U.S. tech firms quickly became an essential part of Mistral AI growth. In January 2023, Mistral AI announced a major commercial agreement with Microsoft, sparking debates about balancing European AI sovereignty with the pragmatic need for American cloud infrastructure and resources.

Defending this partnership, Cédric O argued that collaborating with U.S. companies is simply inevitable and necessary: “I think you have to go back to the origins of the project...At the beginning of Mistral there are three AI researchers...And they wanted to create a company for two reasons. The first is that they said they had the skills to create a company and build a great company in Europe. And the second is that they weren't extremely delighted with the evolution of the technology.”

Given Mistral’s small size and the massive scale of American AI leaders, Cédric O believes partnerships are not a betrayal of Europe’s tech ambitions, but simply a reality of competing successfully in a global landscape. However, he still sees Mistral’s European roots and perspective as core differentiators that guide its ethical AI approach.

The Cybersecurity Risks and Potential of Generative AI

One of the most thought-provoking parts of Cédric O’s keynote was his analysis of how generative AI could impact cybersecurity threats. Responding to a question about this issue, Cédric O argued that in the short term, capabilities like ChatGPT will not radically change the nature of highly advanced cyberattacks. However, generative AI does lower the barriers to entry for basic hacking and phishing.

As Cédric O highlighted, “What I think fundamentally changes generative AI in cyber is not so much hyper-sophisticated attacks, I think it can change things, but really more in the long run. On the other hand, it completely reduces the cost of entry into being able to make malware or being able to make a complete computer attack in a completely practical way.”

By making it easy for anyone to quickly generate malicious code and websites, tools like ChatGPT could lead to a democratization of cyberattack capabilities. While this may not enable truly cutting edge nation-state cyber operations, it dramatically expands the pool of potential threat actors. Companies and governments will need to brace for this new era of “mass-market cybercrime” enabled by generative AI’s exponential improvements in software productivity.

Mistral’s Focus on Developing Core AI Models

Towards the end of his appearance, Cédric O also directly addressed a question about how Mistral will help French companies develop their AI talents and capabilities. In response, he bluntly declared “no, we are not going to support increasing the skills and knowledge of companies.”

As Cédric O elaborated, given its small size, Mistral’s focus is solely on advancing its core AI models, not building an end-to-end business platform. Mistral aims to partner with integrators and other companies that can package its technology into real-world applications.

Cédric O believes this tight focus is crucial given Mistral’s massive disadvantages in resources compared to tech giants like Google and Microsoft. If Mistral tries to stretch itself too thin, it risks losing its most valuable asset – its world-class team of AI researchers. By maintaining ruthless priorities, Mistral hopes to continue pushing the capabilities of AI itself forward in pioneering new ways.

The Need for Open and Honest Reflection

Stepping back from the details, I think the overarching takeaway from Cédric O’s insightful keynote is the need for open and honest reflection on the risks alongside the rewards of progress in artificial intelligence. Cédric O brings the candid perspective of both a startup founder seeking to advance AI capabilities and a former public sector leader charged with steering its safe development.

He thoughtfully wrestles with difficult questions about European autonomy versus American partnership, generative AI’s dual edges of productivity and cyber risk, and startups’ need to balance ambition with disciplined focus. Cédric O neither over-hypes AI nor dismisses its challenges, instead aiming for a clear-eyed view. This willingness to tackle hard truths is exactly what’s needed from all leaders and thinkers navigating our age of intelligent machines.