CIBC Tech & Innovation Market Update – November 2024

Calendar Q3 was all about AI for the big tech names, with capex spending and AI monetization highly topical. Despite headline beats across all the names we track, share price reactions were largely mixed. Google and Amazon performed the best the day following reporting, while Microsoft, Apple and Meta lagged. Similar to last quarter, earnings beats are no longer enough for big tech investors, and investors are clearly looking for demand to support investment when it comes to AI spending. While capex spend is still heavily outpacing AI revenue across the space, several of the big tech companies are now able to quantify the monetization of their AI solutions and offerings.

Facebook and Instagram will once again be using facial recognition technology on photos uploaded to each platform, just three years after Meta shut down the feature following privacy concerns (and a hefty class action settlement). The company says it plans to use facial recognition technology to protect users from scam websites that rely on public figure impersonation and to help users regain access to locked accounts. Whether or not the new use cases are viewed as a violation of privacy acts remains to be seen, but as cybercriminals become more sophisticated and gain access to better tools for identify theft, it begs the question: is providing more personal data the best way to protect personal information?

If you’re an average consumer, you’ve likely paid for a subscription after forgetting to cancel a free trial or paid for an entire extra month after not cancelling a subscription early enough. These extra fees and the hoops to jump through to cancel subscriptions can leave customers feeling frustrated and cheated. However, the days of paying extra fees appear to be coming to an end for American consumers. After several lawsuits targeting specific companies and tactics, the Federal Trade Commission has now introduced a final “click-to-cancel” rule that will require companies selling subscription-based products to make cancelling easier for users. The rule applies to auto-renew subscriptions and includes everything from streaming services to gym memberships. Click-to-cancel is set to come into effect in six months.

A year ago, we were writing about AI hallucinations and the risks that came with relying on GenAI for specialized knowledge. While these risks certainly haven’t disappeared, they may be becoming less of a concern as foundation models become more advanced. Many individuals agree that ChatGPT has improved in terms of the quality of its responses compared to the November 2022 version. One of the reasons for that improvement may surprise you. Since their inception, model builders have been employing thousands of human contractors to label training data and help models sort fact from fiction in an effort to limit hallucinations. As the models have mastered basic facts, more sophisticated training data is needed, which in turn requires more sophisticated labellers and trainers. Historians, doctors, and PhD holders are among some of the top talent now being recruited to train AI models.

California Governor Gavin Newsom vetoed an artificial intelligence bill that would have required AI developers to comply with certain safety rules aimed at curtailing AI misuse. In the absence of federal AI legislation, California’s bill was an attempt to lay the foundation for regulating the technology in the US. Policymakers around the globe, mainly the EU, have taken steps to regulate the rapidly advancing technology but regulation efforts in the US continue to hit roadblocks. With the country divided on how to balance regulation and innovation, the controversial bill has been a particularly hot-button issue. Governor Newsom, a Democrat, had faced pressure to veto the bill from many influential tech executives, while simultaneously being pressured by academics and even celebrities to sign it into law. Ultimately, the Governor stated that while he saw the bill as well-intentioned, he cited the breadth of the bill and the burden it would place on AI companies as well as its threat to California’s position as a leader in the AI space as factors supporting his veto decision.