3 steps to AI mass adoption

As I wrote previously , no matter how much the media write about AI tools implementation in business, replacing jobs etc…we are not yet at mass adoption.

As per August 2024 Statista report, 18% of people between 18-24 years old used chatGPT weekly. And it’s the maximum usage occurrence across all population segments.

I also wrote, and still believe mass adoption will indeed happen but in a timeframe of around 5 years.

Why not before? Why not after? You may ask.

It’s a bit of a long intricate story with few moving parts but I will simplify at least as I see it.

To me, there are 3 major steps in order to see mass adoption of generative AI tools, such as chatGPT.

Step 1: somehow generativeAI tool(s) reaches a ‘stable’ , ‘mature’ stage. This means the underpinning technology does not have any new breakthrough. The results still can improve but with linear increments not massive jumps. To date, broadly speaking LLM transformers for text and Diffusion for images are reaching their full potential. If no new breakthrough is found, we are going to ‘stabilise’ the technology.

Step 2: Once the technology is stable, now you can regulate. See, you cannot regulate a technology that has potential massive improvements because otherwise your rules may be not useful or detrimental : in other words you need to know the limits of a technology before deciding the rules around it. EU AI Act is a first attempt at regulation and it is rolling out right now. Same for company certifications on usage of AI like ISO 42001.

Step 3: Once the rules are clear, now you can insure the services and products. In order for a product or service to be widely used you should know if something goes wrong somehow there is coverage and somehow will pay for the damage, i.e. you want insurance. Insurance can come into place only if the rules are clear (the rules will determine what is the ‘maximum’ legal damage hence you calculate the insurance risk).

I stress these steps are always the case when you want mass adoption. Think about self-driving cars: first you want a ‘stable’ technology (at level 4 or 5 of autonomy).? Then stable rules can be agreed on usage and finally you can have insurance on such cars.

I also stress that in the case of generative AI tools mass adoption will happen in less than 5 years only if no new breakthrough is coming.?

If that happens… who knows what the timeline is…

I hope it helps.

#ai #artificialintelligence #innovation #business #work #technology