Big Tech did no such thing, and AI development continued apace. Realizing his public awareness campaign wasn\u2019t enough, Bengio set about building a product that could curb AI\u2019s worst impulses. In the spring of 2024, after decades of academic research, he began working on what would become LawZero, a research organization dedicated to producing safe AI. The team aims to create an AI model that is predictive but disinterested, explanatory but not manipulative. In doing so, LawZero hopes to save humanity from its own creation.\u00a0<\/p>\n
\u201cThe smarter they get\u2014and they\u2019re getting smarter every month\u2014the worse the safety situation is getting,\u201d said Sam Ramadori, LawZero\u2019s co-president and executive director. He\u2019s already seen teenagers pushed to suicide and hackers empowered by AI models, and he floats the possibility of AI being used for taking down electricity grids, playing havoc with the banking system, bio-weapon attacks and other dire scenarios.<\/p>\n
\t\t\t\t\tRelated Articles<\/p>\n
\t \t LawZero\u2019s secret sauce is what Bengio calls Scientist AI. Whereas commercial AI models imitate human-like behaviour for what critics say is the goal of increasing user engagement, Scientist AI\u2019s model is built on demonstrably true and false statements\u2014humans breathe oxygen or the earth is flat, respectively\u2014making it \u201cepistemically honest,\u201d Bengio and others argued in a research paper<\/a> published in late June.\u00a0<\/p>\n As such, Scientist AI isn\u2019t prone to sycophancy, persuasion and other AI model tendencies that can distort reality and reinforce user beliefs and plans, regardless of how fantastical or dangerous they may be. Nor does Scientist AI have a desire to protect its own operational status, the \u201cself-preservation\u201d tendency<\/a> that has led some commercial AI models to disable their own shutdown protocols, or engage in deception and blackmail to avoid being deleted. If commercial chatbots are the AI equivalent of a bad friend willing to indulge your delusions and worst instincts, Scientist AI aims to be that friend\u2019s nerdy neighbour, there to tell you what\u2019s true and what isn\u2019t.<\/p>\n \u201cYoshua basically said he could invent an AI model in which the core is designed differently to current large language models, so we don\u2019t get the unpredictable and at times dangerous behaviour that we\u2019re getting out of today\u2019s advanced models,\u201d Ramadori said.<\/p>\n Building Scientist AI and developing sellable products from it happens in LawZero\u2019s perch on the second floor of Mila\u2019s headquarters in Montreal\u2019s Mile-Ex neighbourhood. LawZero is a spinoff of Montreal AI research institute Mila, and though it is a non-profit, there are startup vibes to the company. The office is wide open, laptop-strewn and dominated by whiteboards. It has an aspirational tagline\u2014\u201cProtecting human joy and endeavour\u201d\u2014and a cadre of blue-chip backers including the Gates Foundation and the Silicon Valley Community Foundation.<\/p>\n \t\t \tLawZero is a non-profit, but its Montreal offices look like those of a start-up. Photo: Roger Lemoyne for The Logic\t<\/p>\n Ramadori can see LawZero more than doubling the size of its staff, from roughly 40 to 100, within the next year, and to well over 300 in the years to come. In Ramadori, the former CEO of Montreal-based building management startup Brainbox AI, LawZero has an experienced C-level executive who knows how to fundraise and make money.<\/p>\n The company also has a startup\u2019s ambiguity as to what, exactly, the eventual sellable product will be. In the short term, LawZero won\u2019t be OpenAI, which built its own language model from scratch and commanded a US$852-billion valuation as of March.<\/p>\n \u201cAt this point we have multiple pathways that we\u2019re seeing on how our solution will be used at the end,\u201d said Catherine Saine, LawZero\u2019s vice-president of development and operations. The immediate goal, at least for now, is to build a tool that could be affixed to existing models, acting as a guardrail against commercial AI\u2019s worst tendencies.\u00a0<\/p>\n \u201cIf everyone else\u2019s model gets so safe that our product is only 10 per cent better and no one buys it, we\u2019ll still be opening champagne bottles in this place.\u201d<\/p>\n Saine, who worked at Mila before heading over to LawZero late last year, envisions LawZero teaming up with companies, including Cohere, to integrate LawZero\u2019s core methodology into what these companies have built. \u201cMaybe at the end we\u2019ll have a totally new model that could be an alternative to the ChatGPTs or Claudes of this world,\u201d Saine said.<\/p>\n Ramadori and Saine demur when asked about the potential financial benefits of Scientist AI, framing it instead as a world-saving initiative. \u201cIf everyone else\u2019s model gets so safe that our product is only 10 per cent better and no one buys it, we\u2019ll still be opening champagne bottles in this place. We\u2019ve won,\u201d said Ramadori.\u00a0<\/p>\n Still, the market potential of LawZero and other AI trust, risk and security management products is predicted to grow nearly 500 per cent, to US$21 billion worldwide by 2035, according to a report by Precedence Research, a market research firm with offices in Canada and India.<\/p>\n This is in part because the recognized foibles of the technology\u2014hallucinations, flawed reasoning and blackmail among them\u2014are very real insurance liabilities for corporations. Major insurance companies have balked<\/a> at covering AI-related losses. \u201cThe rapid diffusion of agentic AI has created a new coverage problem for commercial insurance,\u201d declared a recent paper<\/a> from insurtech and cybersecurity group AIFT.<\/p>\n \t\t \tLawZero aims to more than double its headcount within the next year. Photo: Roger Lemoyne for The Logic\t<\/p>\n The black-box nature of widely used AI models has hampered end-to-end AI adoption within companies because of the potentially catastrophic dangers of the technology, said KPMG Canada AI research head Andrew Nathaniel Forde.\u00a0<\/p>\n \u201cIs there a trust gap? Yes. And I would go further: it is widening,\u201d Forde said. \u201cConsider why none of us thinks twice before driving over a bridge. It is not that we trust the concrete. It is that we trust the system behind the concrete. An engineer stamped the drawings, an inspection regime keeps it honest, a professional body can end a career over a bad calculation. AI has no equivalent yet. So, people use these systems and withhold their trust at the same time, and frankly, that is the rational response.\u201d<\/p>\n Investment in AI trust and security could effectively pay for itself, as it reassures people and companies as to the trustworthiness of the technology, further spurring the adoption of the technology, according to a March Rand Corporation report.<\/p>\n LawZero has a notable ally in the federal government. During the June rollout of Canada\u2019s AI strategy, AI Minister Evan Solomon said <\/a>the company \u201cis building a new generation of AI intended to provide oversight for the agentic systems being deployed by frontier labs worldwide.\u201d<\/p>\n
\n\t![\"\"](\"https:\/\/www.europesays.com\/canada\/wp-content\/uploads\/2026\/07\/Yoshua_Bengio-Roger_Lemoyne_The-Logic-0U7A6045-1920x1280-1-768x512.jpeg\")
\t<\/a><\/p>\n
\n\t![\"\"](\"https:\/\/www.europesays.com\/canada\/wp-content\/uploads\/2026\/07\/Montreal-April_2025-Mario_Beauregard-CP216409797-1920x1280-1-768x512.jpeg\")
\t<\/a><\/p>\n![\"A](\"https:\/\/www.europesays.com\/canada\/wp-content\/uploads\/2026\/07\/LawZero_offices_Montreal_May2026_roger_Lemoyne10_the_Logic_1920x1280.jpg\")
\t<\/p>\n![\"A](\"https:\/\/www.europesays.com\/canada\/wp-content\/uploads\/2026\/07\/LawZero_staff_Montreal_May2026_Roger_Lemoyne_12_the_logic_1920x1280.jpg\")
\t<\/p>\n