Sanofi CEO Paul Hudson revealed at a panel in Davos on Tuesday that his pharmaceutical company now relies on artificial intelligence to make critical drug development decisions, specifically whether medications should advance to the next phase of development. Hudson’s rationale for incorporating AI into these high-stakes decisions is striking: the technology has no personal or professional attachments to influence its judgment.
Speaking alongside other industry leaders including Amazon Web Services CEO Matt Garman and Aramco CEO Amin Nasser, Hudson explained that when Sanofi’s senior decision-makers convene to discuss a drug’s future, they begin with an AI recommendation. The pharmaceutical giant uses AI to determine whether drugs should “pass through a tollgate” and receive approval to move forward in the development pipeline.
The AI advantage, according to Hudson, lies in its objectivity. “The agent doesn’t have a career at stake,” he noted. “The agent isn’t wedded to the project for the last 10 years. The agent is dispassionately saying: ‘Don’t go forward or go forward faster, or go forward and remember these things.’” This represents a fundamental shift in how pharmaceutical companies approach drug development decisions, traditionally dominated by human executives with personal stakes in project outcomes.
Sanofi’s AI timeline is significant: the company has been practically using AI for approximately three years. Given that drug development typically takes 12 to 15 years from discovery to market, AI has been involved in roughly one-third of the discovery process for some current drugs. The discovery phase, when manufacturers identify compounds that could become new medicines, represents a massive investment—Sanofi spends approximately €3 billion ($3.1 billion) on discovery within that timeframe.
The pharmaceutical company, known for producing Lantus insulin injections and Plavix blood thinners, is betting heavily on AI to streamline and improve its drug development process. Hudson’s comments also addressed widespread concerns about AI-driven job displacement, arguing that “the jobs that are at risk are the jobs where the human isn’t interested in AI. AI doesn’t beat human plus AI.” This suggests a future where AI augments rather than replaces human decision-making in pharmaceutical development.
Key Quotes
And we do that because it’s very sobering, because the agent doesn’t have a career at stake. The agent isn’t wedded to the project for the last 10 years. The agent is dispassionately saying: ‘Don’t go forward or go forward faster, or go forward and remember these things.’
Sanofi CEO Paul Hudson explained why the pharmaceutical company begins drug approval discussions with AI recommendations. This statement highlights the key advantage of AI in decision-making: its complete objectivity and freedom from the personal and professional biases that can influence human executives.
And we’re not used to having somebody without a career at stake in the room at a senior level.
Hudson acknowledged the cultural shift required to accept AI as a decision-making participant in executive-level discussions. This quote underscores how revolutionary it is to have an entity making recommendations without personal consequences tied to the outcome.
The jobs that are at risk are the jobs where the human isn’t interested in AI. AI doesn’t beat human plus AI.
Addressing concerns about AI-driven job displacement, Hudson argued that the real threat isn’t AI itself but rather workers who refuse to adapt and integrate AI into their workflows. This reflects a broader industry perspective on AI as an augmentation tool rather than a replacement.
Our Take
Hudson’s candid admission that AI lacks “a career at stake” reveals both the promise and the uncomfortable truth about human decision-making in corporate environments. While we celebrate human judgment and experience, we rarely acknowledge how self-interest and career preservation can distort critical decisions—especially when billions of dollars and years of work hang in the balance. Sanofi’s approach represents a pragmatic middle ground: AI provides the objective recommendation, but humans retain ultimate authority. This hybrid model may become the template for AI integration across industries where decisions carry enormous financial and human consequences. The €3 billion investment in discovery that Hudson mentioned makes clear why objectivity matters—pharmaceutical companies need to kill failing projects early rather than throwing good money after bad. What’s particularly noteworthy is the timeline: with only three years of AI use, Sanofi is already trusting it for critical decisions, suggesting the technology has proven its value rapidly.
Why This Matters
This development represents a watershed moment for AI in pharmaceutical development, one of the most heavily regulated and high-stakes industries globally. Sanofi’s integration of AI into critical go/no-go decisions for drug development signals that artificial intelligence has moved beyond experimental applications to become a trusted decision-making partner in life-or-death matters.
The implications extend far beyond pharmaceuticals. If AI can be trusted to make billion-dollar decisions about drug development—where human lives hang in the balance—it sets a precedent for AI adoption in other industries facing similar high-stakes choices. Hudson’s emphasis on AI’s lack of “career stake” highlights a crucial advantage: freedom from cognitive biases like sunk cost fallacy, personal attachment, and career preservation that can cloud human judgment.
For the pharmaceutical industry specifically, this could accelerate drug development timelines and reduce the notorious failure rate of drug candidates. With AI providing objective assessments earlier in the process, companies may avoid investing years and billions in drugs unlikely to succeed. This efficiency could ultimately translate to faster access to life-saving medications and reduced healthcare costs, making this a story with profound implications for patients, investors, and the broader healthcare ecosystem.
Recommended Reading
For those interested in learning more about artificial intelligence, machine learning, and effective AI communication, here are some excellent resources: