Mark and the Magic Bean Stock
Liberals plan AI investment despite all rational warnings
In late May, Uber President and chief operating officer Andrew Macdonald said on a podcast that the company had burned through their entire AI budget in four months. Moreover, as Macdonald put it, there wasn't a connection between their AI usage and benefits for their product, making its use "harder to justify." In another case, as Axios reports, one company spent $500 million on Anthropic AI models by accident when the AI company changed their payment model. Oops!
Microsoft's GitHub Copilot, another AI service, also recently moved customers to "token-based billing." This changed a previous $39 a month subscription into a charge-by-use model. As prominent AI critic and PR rep Ed Zitron wrote in his newsletter Where's Your Ed At?,
One burned through 50% of their monthly credits in a single prompt, another burned 60% in the space of a few hours, another 31% in a single prompt, another estimated that they’d burn their monthly credits in the space of a single five hour session, another burned nearly half of their credits in eight prompts, another around 14% of their credits in two prompts, and another lamented that GitHub Copilot had gone from their favorite subscription to their most-stressful overnight after burning 33% of their monthly balance in a few hours.
OpenAI CEO and alleged sexual abuser Sam Altman confirmed these fears, stating that AI costs are "a huge issue," with one top spender using 100 billion tokens a month. Meanwhile, data centre investments appear to be mostly speculative. These are bad signs for the feasibility of AI. Indeed, billionaire investor Ray Dalio of Bridgewater Associates, one of the world's biggest hedge funds, warned that the AI boom is looking increasingly like a stock market bubble.
Anyway, CBC News reported earlier this week the Liberal's draft AI strategy intends to boost business adoption of AI from 12 per cent to over 50 per cent by 2030, support "creating up to 250,000 new jobs" through AI adoption and "up to 90,000 AI-related jobs and work opportunities for young Canadians."
What will this cost? Here are the numbers pulled from an informative BetaKit article. The AI Compute Access Fund, a resource for businesses with compute costs between $100,00 to $5 million, will receive $700 million. Further, a $500 million Canadian Tech Growth Fund will be established, and the Regional Artificial Intelligence Initiative will receive $500 million.
All in all, this includes $2.3 billion in government funding, primarily as investment in private business.
Scott Martin
This comes as the societal costs of AI are becoming harder and harder to ignore. Besides OpenAI's neglect of reporting safety issues that could have prevented the tragic shooting in Tumbler Ridge, BC, the company is also being sued by the state of Florida, in which a shooter allegedly used their product ChatGPT to plan their mass violence. In the social media sphere, Meta's AI models allowed hackers to access popular accounts by just asking it. Google employees are mocking the company's AI effectiveness while the CEO claims 75% of the code is AI-generated.
Meanwhile, while Carney promises every postsecondary student access to an AI agent, data is coming out that shows a relation between increased AI use and a loss in skills knowledge.
As I've written before, journalism as an industry has especially been negatively affected by the adoption of AI. When the Toronto Star investigated the use of "strong mayor powers" throughout Ontario, they used AI. It was delayed a month as a result. To be clear, AI models, verified by human actors, were used in the investigation, but as Star Editor-in-chief Nicole MacIntyre wrote, "I understand that how we approach AI will help determine whether that trust is strengthened or weakened. I feel confident we can get this right."
Last year, Nicholas Hune-Brown at The Local exposed a freelancer who likely used AI to get published in publications like The Guardian. Toronto local publication The Grind delayed an issue due to similar problems.
In the AI announcement package, much hay is made about protecting Canadians from data breaches, deepfakes and keeping privacy. However, little detail is provided on how this will be achieved.
Despite all of these issues, AI is still seen a silver bullet by banker-brained economists. While Canada languishes in an official recession, Correctional Services Canada is looking into AI to help with transparency. Minister of Artificial Intelligence and Digital Innovation Evan Solomon confirmed that Canada was part of an AI program by Anthropic, which increased token costs by 113 per cent in April. Though, at the time, this was from six dollars to 13, with the further costs of tokens explained earlier, this entwinement does not bode well.
However, looking at AI investment as a benefit to Canadians and innovation is the wrong way to look at it. Initiatives and rhetoric such as this provides a simple and crucial narrative for technocrats like Carney: It is good for the "the economy" (read: the stock market). There is indication that AI investment is the primary reason the US stock market has seen growth, despite little tangible returns on investment. It is a speculative growth at this point, to be sure, but all speculative growth does not indicate a bubble. However, growth with as little benefits as AI indicates more of a bubble than just the numbers of its speculation. On the widespread adaptation front, there remain acceptance issues. While there is no direct example of the Canadian public's view of AI, the US public despises it.
It doesn't matter. Since AI provides boosts in the stock market, the only measure of the economy that neoliberal technocrats like Carney care about, it will be instrumental to their spending plan. Lip service will be paid to the risks of this implementation, but the real priority remains widespread adoption at any cost... to you of course. The investor class will get all the benefits.
