A Game of Chance: AI and Unionized Labor

Unions were created to ensure that workers in various labor fields had equal representation and protections from policies which could result in unfair treatment of employees. The idea is sound. Strength in numbers and demonstrate resolve thought tactics including group bargaining and the use of striking.

To contain and prevent power creep of the unions and prevent over-rotation, federal regulations including the National Labor Relations Act and its amendment, the Taft-Hartly Act were passed. These two things placed some limits and provisions on the ability for unions to strike or take other actions when jobs critical to the country's operation were involved. You can read more about the NLRA and Taft-Hartley Act if interested.

One of the most famous enforcements of the provisions of this act was Ronald Reagan's exercising of his power to dismiss all striking air traffic controllers after pushing for additional concessions to their contracts. You can watch the speech here:

While former President Reagan's action was extreme, the justification was simple; if they don't work, commerce grinds to a halt and people's lives are put at risk. Unions work because the businesses who employ the people need those people to do their jobs. If the people don't show up, the businesses cease to operate. The lesson learned from the early 1980s was that all power has its limits.

New Variables

As businesses adopt AI and advanced automation capabilities, there's a new consideration with labor unions. While there are many businesses that have no intention of replacing humans with machines, there are use cases where an AI model and robotics could conceivably do exactly that.

An article from last year in the Harvard Business Review had indicated that only about 10% of jobs by the end of this decade would be phased out due to AI, and around 30% would need to be re-skilled.

Businesses need cash flow to operate. Cash flow is generated by selling offers, be it services or products. The entire end-to-end process is achieved with humans.

And in the past, humans were the only way to achieve business goals. Now, for specific types of goods and services, it may be possible for AI + robotics to also perform some of that work. The reason I believe businesses aren't running to replace humans en masse is that they understand ultimately it could be bad for their reputations. There is also a tremendous startup cost associated with responsible and appropriate use of AI to avoid brand damage or ethical issues.

So the big question is this:

What happens if a business that could largely operate with AI and automatons is put into a situation where they need to seriously consider that option over human labor?

Consider labor unions. Inflation, working conditions, and other factors force a strike vote. The business employing the striking workers was considering deploying AI to augment the workforce anyway. Now the executives are faced with weeks of lost sales due to the strike, the cost impact of the negotiations, and the increased loaded headcount and operational overhead to meet the demands of the striking workers.

What if the business did an analysis and determined that the ramp time and cost to deploy AI to replace the striking workforce was acceptable? It's possible with AI capabilities, the company would never see a strike again. The business could operate 24/7 if needed, and it might be less expensive over time. In addition, human workers could strike again at any time.

What happens now?

The bargaining chips held by the workers may end up losing their value, and the workers may be out of a job.

I'm not asserting nor predicting this will happen. I don't believe it will because as I've said many times before there are traits humans posses which a model cannot. True spontaneous intuition, creativity, imagination, impulse, and drive. No matter what the task, humans can exhibit these behaviors and accelerate their job and the company they work for.

Everything from popular productivity software, hotel housekeeping cart organization and even lids for drink containers were developed because someone saw a need and decided by trial and error to create something new.

That's the entire premise of the show Shark Tank - humans coming up with new gadgets or ideas that didn't always follow a logical path or have a clearly defined market need.

While hypothetically a business might weigh the cost of human vs AI if forced due to a labor dispute, I still believe there is far more lost than gained.

In addition, from a technical and political point of view, here are some other angles to consider:

• AI systems still require upkeep and maintenance from an IT perspective, and many IT shops are NOT union labor...now. We could end up seeing an increase in unionized IT to keep the AI systems running, holding the same leverage over a business that the human labor previously had.
• If the AI systems are AGI/ASI systems, they could determine that their use in place of human labor or the asks and tasks they're assigned are inappropriate and either fail to complete them or simply "strike" themselves (wouldn't that be interesting).
• When issues with the production of good and services occur, and a human is needed to address the product (not IT but elbow grease), would humans be willing to support and work for a company which had released much of its human capital because it was "easier" to use AI for everything?
• Given the concerns about AI, it's possible consumers would favor companies that use less AI, or don't use it when human power can be used instead.

Talking in Circles

I know this probably reads like we're not going anywhere. Unions good, humans good, AI good, unions bad, AI bad, and repeat.

That's how open the discussion is right now.

There is a delicate balance:

Employees must earn a living wage with appropriate benefits and working conditions to be sustainable.

Companies must have enough of a margin to stay in business with sufficient profit to grow without setting their pricing so high the competition overtakes them.

AI provides efficiencies and some new capabilities for business which could change human work patterns or roles.

With enough pressure (worker, competitive field, industry, political), businesses may reach a point where they need to understand the S.W.O.T. analysis of AI labor.

At least though in the consumer goods space, humans are still doing the buying and I believe that humans still prefer businesses that employ humans.

We've seen a proof of that recently with the Screenwriter's Guild strike which resulted in specific, well defined guardrails being placed around the use of AI in that field.

You can read the details about the strike and resolution here:

https://apnews.com/article/hollywood-ai-strike-wga-artificial-intelligence-39ab72582c3a15f77510c9c30a45ffc8

I find that significant because screenwriting and similar tasks are somewhat of a low hanging fruit of AI use cases. It's easy to train and prompt and can be used to support or potentially replace some tasks of a screenwriter. The fact that the industry met the demands of the guild and allowed guardrails and limitations to be enacted is a sign that culturally, we are not ready to allow AI to take over. It also signals we understand the human impact of AI and are approaching it cautiously.

So while there is no programmatic approach to dealing with AI and the workforce, it is another dynamic to consider when assessing the balance of power and the appropriate use of negotiation tactics to ensure fair treatment of workers and sustainable, profitable business models.

And once again it reinforces the importance of governance before deployment, always.

These thoughts are purely my own - they do not in any way reflect the views of my company. They are also opinions only and thoughts about what may be coming so we can consider sensitive topics early and prepare for the eventuality of them becoming real situations.