Canada's 'AI for All' Strategy Criticized for Secret Palantir Contracts and Lack of Transparency

Summary

Canada's federal government announced its 'AI for All' strategy this month, positioning itself as a 'strategic anchor customer' to boost domestic AI adoption from 12% to 60% by 2034. However, a critical opinion piece reveals a major contradiction: the government already purchases advanced AI systems—including Palantir's data-fusion and decision-support platforms—but does so in secrecy rather than transparently supporting Canadian vendors. The Department of National Defence signed a contract with Palantir's Canadian arm in March 2020 for $14.4 million, which quietly grew to $44.4 million by October 2025, with approximately $46.8 million actually spent. The Ontario Provincial Police have been running Palantir's Gotham platform since 2015, examples of precisely the sovereign systems the new strategy claims Canada must develop domestically.

The author argues the strategy's approach undermines its stated goals. Rather than direct procurement from Canadian firms, the 'AI for All' plan emphasizes equity stakes in promising businesses, compute infrastructure funding, and certification programs—mechanisms that avoid the transparency and speed of simple purchasing contracts. The piece contends that a 'strategic anchor customer' should buy openly from Canadian vendors under public rules, not accumulate equity stakes that turn companies into 'wards of the state,' and questions why health—the slowest-procuring, most risk-averse sector—was chosen as the proving ground instead of courts, public safety, or defense.

• Choosing health as the missions program pilot risks stalling in pilots for years; defense and public safety would be higher-consequence proving grounds