Speech (committee meeting)

{
    "time": "2026-02-02 16:00:00",
    "attribution": {
        "en": "David Krueger",
        "fr": "David Krueger"
    },
    "content": {
        "en": "<p data-HoCid=\"9156965\" data-originallang=\"en\"> If I understand the question, it's how we would get government to a place where it has a similar level of technical competence with AI as compared to the AI companies. I think that's very difficult. The first step is probably to try to hold the companies more accountable and to regulate them more seriously.</p>\n<p data-HoCid=\"9156966\" data-originallang=\"en\">One reason I think they have so much more talent is they're able to pay much higher salaries because they're getting a lot of investment based on this premise that they will build AGI and superintelligence and then, again, take everybody's job and make trillions of dollars from that. Right now, there's a huge imbalance in terms of the talent, and that's not something that can easily be corrected, frankly. It would be a pretty lengthy process to fully address that.</p>\n<p data-HoCid=\"9156967\" data-originallang=\"en\">On the other hand, one of the main issues right now is a lack of transparency into AI companies. There's very little oversight in terms of what they're doing. It used to be the case maybe five years ago that these companies would publish most of the research they did, so there was a shared understanding in the research community of how the most advanced systems work. That also helped audit the systems for safety more effectively. This has stopped being the case. Companies really don't release very much information at all about their systems, and there's no independent oversight that is mandated.</p>\n<p data-HoCid=\"9156968\" data-originallang=\"en\">We need programs for independent experts to be able to look at the details of the AI systems. That doesn't just include evaluating or testing the model, but also includes having the necessary information about how the model was developed and what data was used, which is often copyrighted data. There have been a number of large-scale lawsuits about that. Also, what training methods were used? Was it trained in a way that is known to cause addiction, dependency and other psychological issues? We know this is something that has happened repeatedly in the tech industry with social media and with AI. Finally, how is the system being deployed and with what additional safeguards and guardrails? What is the company doing to monitor how it's being used? Companies have a strong incentive to understand how their products are being used in order to make more money from them.</p>\n<p data-HoCid=\"9156969\" data-originallang=\"en\">That's something the government also needs to\u2014</p>",
        "fr": "<p data-HoCid=\"9156965\" data-originallang=\"en\">Si je comprends bien la question, il s'agit de savoir comment nous pourrions faire en sorte que le gouvernement poss\u00e8de des comp\u00e9tences techniques en mati\u00e8re d'IA comparables \u00e0 celles des entreprises du domaine. \u00c0 mon avis, c'est tr\u00e8s difficile. La premi\u00e8re \u00e9tape serait plut\u00f4t d'essayer de responsabiliser davantage les entreprises et de les r\u00e9glementer plus s\u00e9rieusement.</p>\n<p data-HoCid=\"9156966\" data-originallang=\"en\">Selon moi, si ces entreprises attirent autant de talents, c'est notamment parce qu'elles peuvent offrir des salaires beaucoup plus \u00e9lev\u00e9s gr\u00e2ce aux investissements consid\u00e9rables qu'elles re\u00e7oivent, dans l'espoir de cr\u00e9er une intelligence artificielle g\u00e9n\u00e9rale et une superintelligence pour ensuite, l\u00e0 encore, remplacer les emplois de tout le monde et en tirer des billions de dollars. \u00c0 l'heure actuelle, le d\u00e9s\u00e9quilibre en mati\u00e8re de talents est immense et, en toute franchise, il ne se r\u00e9sorbera pas facilement. Il faudrait un processus assez long pour y rem\u00e9dier compl\u00e8tement.</p>\n<p data-HoCid=\"9156967\" data-originallang=\"en\">Par ailleurs, l'un des principaux enjeux en ce moment est le manque de transparence dans les entreprises d'IA. Leurs activit\u00e9s ne font gu\u00e8re l'objet de surveillance. Il y a peut-\u00eatre cinq ans, ces entreprises publiaient la plupart de leurs travaux de recherche. Le milieu de la recherche pouvait ainsi acqu\u00e9rir une compr\u00e9hension commune du fonctionnement des syst\u00e8mes les plus avanc\u00e9s. Cela permettait \u00e9galement une v\u00e9rification plus efficace des syst\u00e8mes de s\u00e9curit\u00e9. Ce n'est plus le cas aujourd'hui. Les entreprises ne publient presque plus rien au sujet de leurs syst\u00e8mes, et il n'y a aucune surveillance ind\u00e9pendante obligatoire.</p>\n<p data-HoCid=\"9156968\" data-originallang=\"en\">Nous avons besoin de programmes permettant \u00e0 des experts ind\u00e9pendants d'examiner les d\u00e9tails des syst\u00e8mes d'IA. Il ne s'agit pas seulement d'\u00e9valuer ou de tester le mod\u00e8le, mais aussi de disposer des renseignements n\u00e9cessaires sur la conception du mod\u00e8le et l'utilisation des donn\u00e9es, lesquelles sont souvent prot\u00e9g\u00e9es par des droits d'auteur. Il y a d'ailleurs eu un certain nombre de poursuites judiciaires d'envergure \u00e0 ce sujet. Il faut \u00e9galement savoir quelles m\u00e9thodes d'entra\u00eenement ont \u00e9t\u00e9 utilis\u00e9es. Le mod\u00e8le a\u2011t\u2011il \u00e9t\u00e9 entra\u00een\u00e9 d'une mani\u00e8re susceptible de provoquer d\u00e9lib\u00e9r\u00e9ment une accoutumance, une d\u00e9pendance et d'autres probl\u00e8mes psychologiques? Nous savons que ces d\u00e9rives se sont produites \u00e0 maintes reprises dans l'industrie de la technologie, par l'entremise des m\u00e9dias sociaux et de l'IA. Enfin, comment le syst\u00e8me est\u2011il d\u00e9ploy\u00e9, et que faut\u2011il ajouter comme balises et mesures de protection suppl\u00e9mentaires? Comment l'entreprise surveille\u2011t\u2011elle les usages qui en sont faits? Les entreprises ont tout int\u00e9r\u00eat \u00e0 comprendre comment leurs produits sont utilis\u00e9s pour en tirer davantage de profits.</p>\n<p data-HoCid=\"9156969\" data-originallang=\"en\">C'est quelque chose que le gouvernement doit aussi...</p>"
    },
    "url": "/committees/ethics/45-1/25/david-krueger-3/",
    "politician_url": null,
    "politician_membership_url": null,
    "procedural": false,
    "source_id": "13326648",
    "document_url": "/committees/ethics/45-1/25/",
    "related": {
        "document_speeches_url": "/speeches/?document=%2Fcommittees%2Fethics%2F45-1%2F25%2F"
    }
}