January 25, 2024

Now that the "new thing" hype has passed a little, investors in AI are really making some money.

Chip companies and AI-focused tech firms are driving the growth of the stock market. The rate of growth has been compared to the Dot Com bubble for its exuberance. There is a little to that given that the market price increase for Nvidia stock has grown at a similar (faster?) rate than Cisco Systems did in the 1990s before it collapsed by over 90%.

Tech companies are back in the race to nowhere as investors pour money into buying their shares. Many of those companies are trading at around 40 times their value.

The question for companies who are not "tech companies" is what do you invest in production wise?

Maybe companies do not have to worry about it? AI has already passed into a sort of "it is in everything" phase. Ads reference AI in some way and all new products tout some sort of "AI" capability.

It is likely that many of the AI systems that will increase productivity will slowly slip into the workflow of companies. That is already happening with the use of generative AI by workers. But, much of this will not directly impact the number of jobs in a visible way as some workers are duped into taking on some of the costs of implementing efficiencies themselves.

The issue for companies is how you leverage these technologies faster and better than other companies at the process level?

The answer is: new workers with new skills.

The firms that will do better are the firms who hire people who understand and know how to implement new technologies into their production processes. On the flip-side, re-tooling workers through retraining and reallocation of work inside companies will be necessary as well since new techno-aware hires will not understand current processes and managers are usually too quick to expect revolutionary change in productivity levels or benefits to their own metrics.

There are several problems.

• Management are among the jobs that are more likely to be replaced by AI systems.
• Managers are mostly bad at understanding how to implement new technologies because they only understand (maybe not even that) the old technologies.
• Workers do not like being told that they have to gain some skill-sets to keep their jobs.
• Shifts in employment come at a pace that is usually difficult for people in the real world to adapt to.
• Price of debt is still rather high and companies are slow to borrow to implement brand new tech (unless everyone else is and they are forced to).

This all leads to delay which leads to unnecessary layoffs and work environment disruptions later on.

Unions need to identify the potential risks for members in this environment and start talking with their members about just transition in this context. The goal is to negotiate language in their collective agreements that deals with technological change, language that is more robust than current language because in many workplaces the impact is going to be broader than the pace of change in the past 20 years.

New areas of investment for firms is also going to put pressure on workers as spending on new physical capital and new employees with the skills to use that new technology is going to cost a lot.

Alternatively, the development will be done through outside specialist firms who rent technology transition to other firms. This will be much more difficult to address through standard CA language as we have seen in the outsourcing of IT and other tech services in the previous two decades.

The building of upgraded IT systems is a long-term program and usually done with an eye to eliminate workers or replace them with cheaper, less skilled ones. This will be no different. The centralization of the few workers with skills to implement AI and the push to outsource that knowledge is going to be high in medium and smaller firms.

These are all things that the labour movement will have to contend with. And, we are going to have to do it better than we deal with the casualization of work through digital platforms.

As the price of debt comes down, I am concerned that workers and their unions are going to be caught out in an environment of yet another force pushing rapid change to work environments, beyond climate change, demographics, costs of living, and shifting geopolitical conflicts. I am concerned because the regulatory environment that would facilitate a just transition still alludes capitalist countries and there is a sense in political leadership (of all kinds) that things are moving too quickly to regulate.

If there is one thing I have to outline it is that slow regulatory implementation can be a feature, not a bug. Regulation is responses to bad things happening. Good regulation is responsive, not static, and it is based on good data and public pressure and driven by people with high levels of knowledge in the sectors being regulated.

Public pressure is something workers and their unions are good at. Data, not so much. The class has an opportunity to push regulation at a speed that will respond to the worst aspects of this while building consciousness of why it is happening. We must demand investment in open data collection and analysis from our national research institutes. Those institutes were gutted in the 90s in massive cuts that cut everything else, but if we want "smart regulation", we have to employ the smart people who can oversee what is going on and push them on the political economy they need to regulate it properly.

It may not be a sexy call of the left, but it is an essential one to win against the invisible hand trying to undermine the quality of life for workers to pad the pockets of capital.