Company

    A team that turns requirements into proof

    Mankinds audits the behavior of AI systems against the requirements that apply to them, as they actually run, and keeps the evidence a third party can review.

    Talk to an expert
    We are hiring

    Hover over a face for more details.

    • Lancelot, AI Safety Specialist. Before Mankinds: Square Management, Evolem. A born salesman, with a heart. He never drops a follow-up, and he asks how you're doing.
    • Stanislas, Head of AI Policy. Before Mankinds: Orange, Giskard, Xiaomi. Navigates everything: laws, institutions, corridors. He always knows the way, and someone at the end of it.
    • Alexandre, Head of AI. Before Mankinds: Criteo, Dataiku, Inria. Works at any hour. A message at 4 a.m.? Not a bug.
    • Baptiste, CTO & Co-Founder. Before Mankinds: Free, Société Générale, OCBC. A bulldozer. He doesn't work around technical problems: he goes straight through.
    • Laurent, CEO & Co-Founder. Before Mankinds: Free, Veepee, Louis Vuitton. Gadget addict. While the team lifts weights, he lifts preorders.

    Our story

    It all started with a conviction

    Laurent and Baptiste spent years building and securing systems used by millions of people, in telecoms, banking and e-commerce. Dedicated teams, audits, procedures: they know what it takes, and where it stops. They came away with one conviction: declared compliance, without continuous technical verification, protects no one.

    With generative AI, the risk changes scale. Agents act on their own, with access to data and tools, and their behavior shifts with every version. They decided to take the problem head on.

    Around them, a complementary team: Lancelot to take Mankinds to market, Alexandre to stay a step ahead in research, Stanislas to stay at the forefront of the law. All share the same drive: to protect, and to make it safe to build with AI.

    Our values

    What guides us

    • 01Check before we claim

      An audit is only worth its rigor. We test before we conclude, we say where our results stop, and we publish our method so others can challenge it.

    • 02Law and code at the same table

      The risk of an AI system belongs to no single team. Lawyers, researchers and engineers work on every case together, and we bring your legal, product and engineering teams to the same table.

    • 03AI that works for people

      It is in our name. We do not want to slow AI down: we want it deployed safely, for the people who build it, the people who use it and everyone it affects.

    Talk to the team

    A call with an expert: your use case, the requirements that apply to it and how they are tested.

    Talk to an expert