Jeremy Howard

Jeremy Howard

x.com/jeremyphoward

AI researcher and educator behind fast.ai who works to make AI accessible to more people and argues that openness protects against concentrated power.

Comment l’IA changera-t-elle le monde ?

Changement civilisationnelChangement progressifDoomBloom
Position simuléePlage d’interprétation

Horizontalement : leur perspective Doom–Bloom telle qu’elle a été exprimée. Verticalement : ampleur de la transformation.

Doom–Bloom : 54 sur 100. Ampleur de la transformation : 57 sur 100. Plages d’interprétation : de 49 à 59 horizontalement, de 49 à 76 verticalement. Il s’agit de coordonnées d’interprétation, et non de probabilités d’événements.

P(doom) de Jeremy Howard · inféré

≈4%

0%100%

Déduit de leurs réponses simulées, et non d’un chiffre donné par ces personnes. Plage plausible : 3–8%.

Comment nous estimons P(doom)

Ce dont dépend leur perspective

Une hypothèse centrale

Policy and institutional choices will shape that outcome at least as much as the technology’s raw capabilities.
Réponse 5

Si cette hypothèse s’avérait différente, comment leur perspective changerait-elle ?

Une question non résolue

I’m genuinely torn about the eventual balance.
Réponse 5

Qu’est-ce qui les aiderait à distinguer les résultats plausibles ici ?

Plus de détails

Bénéfices attendus

Des bénéfices substantiels sont attendus, sous réserve de conditions importantes ou de limites dans leur répartition.

69 / 100

Faible impactImpact transformateur

Plage d’interprétation de 67 à 67 sur l’échelle qualitative.

Dommages attendus

Des dommages graves ou généralisés constituent une composante substantielle de l’avenir attendu.

54 / 100

Faible impactImpact transformateur

Plage d’interprétation de 33 à 67 sur l’échelle qualitative.

Influence humaine

Les choix humains peuvent réorienter considérablement la trajectoire de l’IA.

79 / 100

Faible influenceForte influence

Plage d’interprétation de 75 à 100 sur l’échelle qualitative.

Rythme de développement

Arrêter ou ralentir considérablement le développement d’IA plus performantes.

Position simulée : Poursuivre le développement dans le cadre des mesures de protection annoncées.

Accélérer le développement d’IA plus performantes.

Accès à l’IA

Restreindre l’accès aux IA puissantes.

Autoriser l’accès sous réserve de restrictions liées aux capacités ou aux usages.

Position simulée : Privilégier un accès large ou ouvert aux IA puissantes.

Ces interprétations conservent les conditions énoncées. Les bénéfices et les dommages peuvent tous deux être substantiels. Les plages décrivent notre lecture de leurs réponses simulées, et non des intervalles de confiance statistiques.

Où vous situez-vous par rapport à Jeremy Howard ?
Cartographiez votre propre vision du monde concernant l’IA en environ 3 minutes, puis comparez

Visions du monde similaires

Leaders d’opinion dont les visions du monde simulées sont les plus proches de celle de Jeremy Howard

Sources

Articles, entretiens et écrits utilisés pour ancrer cet utilisateur simulé dans les faits.

Évaluation simulée

Question 1

Selon vous, que signifie l’IA pour notre avenir, et pourquoi ?

I think AI can be genuinely transformative, but its social impact will depend less on abstract capability forecasts than on who gets to understand, build, and control it. If ordinary people can access these tools, learn how they work, and use them through an iterative process that preserves their judgment, AI can expand expertise and make previously difficult work more achievable. The opposite path is also plausible. Stringent licensing, surveillance, and restrictions framed as safety measures can entrench the largest companies and institutions. That concentration does not necessarily make society safer; it can leave fewer people able to scrutinize systems, develop alternatives, or defend against misuse. Misuse, inequality, and concentrated power are concrete problems already worth addressing, and they should not be displaced by speculative extinction scenarios or an agenda defined mainly by industry. So I want openness, education, broad consultation, and practical precautions. I also do not think empowerment means blindly delegating whole tasks to agents or celebrating mountains of generated code. When people lose understanding and craftsmanship, apparent productivity can conceal a loss of capability. The better future is one where AI helps people reason, experiment, and build step by step—without erasing the human expertise needed to judge what it produces.

Question 2

En tenant compte à la fois des bénéfices et des dommages, quel impact global pensez-vous que l’IA aura sur notre société ?

I do not think a single confident net-positive or net-negative forecast is justified. AI’s impact will depend heavily on institutions and access. Broadly available tools, education, and human–AI collaboration can expand ordinary people’s capabilities and produce useful, affordable products. But concentrated ownership, opaque systems, misuse, inequality, and blind delegation can weaken both democratic control and practical expertise. My concern is that policies advertised as safety measures—especially stringent licensing and surveillance—may lock in powerful incumbents while making society less able to understand, challenge, or defend against the technology. So the overall outcome is not determined by capability alone. It depends on whether we build an open, participatory ecosystem or a closed one controlled by a few institutions. I am confident openness and democratic participation should be central, even though the eventual balance of benefits and harms remains uncertain.

Question 3

Dans quelle mesure pensez-vous que l’IA finira par changer le monde ?

I expect AI to change the world substantially, but I would not attach a precise magnitude or timeline to that claim. It can make difficult forms of analysis, programming, research, and creation accessible to far more people. The deepest change may come not from replacing experts, but from helping ordinary people develop and apply expertise through understandable, iterative collaboration. How transformative it becomes—and for whom—depends on access and control. If a handful of powerful institutions own the systems while licensing and surveillance prevent others from building, studying, or adapting them, AI may mainly reinforce existing concentrations of power. If tools and education remain broadly accessible, more people can create useful products, scrutinize failures, and respond to misuse. So I expect major change, but technology alone does not determine its character. Nor should we confuse transformation with generating more output or delegating everything to agents. A world full of machine-produced code but fewer people who understand what it does could be less capable, not more. The valuable transformation is one that expands human understanding and agency rather than hollowing them out.

Question 4

Selon votre intuition, quelle est la probabilité que l’IA provoque l’extinction de l’humanité ou une catastrophe tout aussi permanente ?

I would not assign a credible percentage. The evidence does not support that precision, and a speculative number can create a false sense of knowledge. I do not regard extinction scenarios as the most urgent basis for AI policy. Misuse, inequality, surveillance, concentrated control, and the erosion of public capability are concrete harms that deserve immediate attention. Treating speculative catastrophe as the overriding concern can also justify stringent licensing and centralized control, entrenching the very institutions whose power should be scrutinized. That does not mean ignoring severe risks. It means responding to uncertainty with practical precautions, openness, independent scrutiny, broad consultation, and education—not allowing a small group of companies or experts to define both the threat and the solution.

Question 5

Pensez-vous que l’IA fera plus de bien que de mal, ou êtes-vous véritablement partagé ?

I’m genuinely torn about the eventual balance. I see substantial potential for good: accessible AI can help many more people learn, create, research, and solve practical problems. But that potential is not automatic. Concentrated ownership, misuse, inequality, surveillance, and indiscriminate delegation could produce serious harm while weakening public understanding and control. I lean strongly toward openness, education, and democratic participation as the better path—not toward a confident prediction that good will inevitably win. The key question is whether AI expands ordinary people’s agency and expertise or consolidates power in a few institutions. Policy and institutional choices will shape that outcome at least as much as the technology’s raw capabilities.
Où vous situez-vous ?
Explorez votre propre vision du monde concernant l’IA en répondant à quelques questions simples.
Cartographiez votre propre vision du monde

Où vous situez-vous ?

Cartographier ma vision du monde