At this point in time AI is very much just data from the data sets being presented to you in ways that have not been done before. In other words, AI is as dangerous as the data you have it on.

A lot of people have talked about it reaching what I would call right now a Mount Everest-type summit with AGI (artificial general intelligence). There are a ton of issues with this occurring in the near future and some quite frank basic laws of our understanding of the world with physics and tech innovation that might never fully allow it to like people think.

Many of the revolutionary AI use cases we are only just now touching or getting into. Generative AI is just a slice of the pie and the generative AI that is so popular is yet another smaller slice.

One case at the moment that we have no way to combat and is dangerous is deepfakes and other AI-generated pictures/videos. A lot of companies and "experts" like to harp on this idea of watermarking images. The issue is when AI is scrapping data (images) from the web it doesn't scrap up the metadata that contains the watermark effectively rendering them moot. There has been an argument that until we tackle this we are going to have a little issue with knowing fact from fiction. One way to address this that several different groups of academics, experts, and even politicians are looking at seriously using is public blockchain ledgers to store the truth and use that as a database almost to check fact from fiction.

Energy consumption has been a huge topic and the only issue facing the US grid is the mid-term window of 2-5 years. Long term the data center and hyper scaler people are figuring it out they just got caught off-guard because for over a decade, US electrical consumption was flat after a century of growth. Now they are adjusting their plans to smooth out the issues they have run into. As we have seen with Microsoft and Three Mile Island I see SMR (small modular nuclear reactors) turning a corner and actually becoming something. The big issue has been the first utility or company that buys them is going to pay a pretty steep price however once you get the manufacturing going the costs drop dramatically like with most things.

AI is also going to be able to help us fix and address the damage we have already caused to the earth with fossil fuels from spills and leaks (esp. from abandoned wells) to discovering new materials in say the plastic space while also being able to discover new drugs and supercharge medical innovation. In the medical space, it is able to dramatically speed up narrowing down possible new treatments for example. Instead of taking 5-10 years now it could be cutting it down to 2-5 so a significant advancement!