Big Tech?
Throughout these blog posts, we have heard how AI is changing industries left and right. From sports to accounting to traffic management. There is no stone that is being left unturned. I want to take a step back and simply look at what makes up Artificial Intelligence. At its simplest, an AI system uses computer science logic (algorithms/Language models/Machine Learning) based on huge amounts of data to solve problems. A theme that we have discussed in this class is that the data is king. We have seen evidence of that from the Target case study where they found their AI hiring algorithm to be biased and sexist. The reason for this was due to the data it was trained or created on. Professor Kane stated in class “An AI system is only as good as the data it is trained on.” Okay, so now let’s shift our focus on the data. ChatGPT for reference, was trained on around 570 GB of datasets. This included pretty much everything you can think of from the internet (books, StackOverflow, wikipedia, etc.) I think we can all agree that ChatGPT has shaken up the world and has brought AI to the spotlight. Where we are on Gartner’s hype cycle is unknown, and where we are in terms of controlling AI is also unknown. Recently we have seen where many prominent people in the tech world petitioned to put a halt on AI development. This comes back to the TedTalk we watched by Sam Harris “Can we build AI without losing control over it.” However, for this blog post, we are going to assume we are not halting. Heck, Musk himself supported the halt and we saw within the past few days his desire to create his own AI, TruthGPT, ironic. So we have to understand that this current hype was created off of only 570 GB of datasets. Google on the other hand, collects 2.5 exabytes (2,500,000,000 GB) of data every single day. Now, we know google Bard isn’t trained on all of that data but it is reasonable to assume that the amount is far greater than 570 GB. Now, I have not personally used Bard but from my understanding, gather that it is pretty similar in functionality to ChatGPT. I want to come back to Sam Harris’s talk. One of the assumptions that he made was that we are going to continue to improve and develop our intelligent machines. If that is true, and if the AI is only as good as its data. It is reasonable to assume that in the near future, Google will win the generative AI race. Some skeptics may point to Microsoft integrating ChatGPT with Bing, however, Google owns 93% market share in the search engine space. It just came out recently that Amazon is going to be joining the fight. Not by creating a chatbot, but by developing two AI language models. What does this look like? I have no idea, but I will be in anticipation to see what develops. To circle back to my original question, who is going to win, it is hard not to say google.
Governments?
So we just went through the generative AI race broadly between mainly Microsoft and Google. However, another aspect of this race will be between governments. A recent global innovation wave was the emergence of Big Tech (Apple, Microsoft, Facebook, Google, Amazon.) On a global scale, it is safe to say that the U.S. won that specific battle due to the fact that these are all American companies so the U.S. reaped the economic rewards. Countries all across the globe understand this and are now attempting to win the AI race. The implications of winning this race, however, are far more severe. Not only for economic and consumer dominance, but for physical and military dominance. Going back to the Sam Harris talk, a 6 month lead in AI development is similar to 50 years of normal development. I know that might sound a little butchered but ya’ll get the point. With such a small margin of error, will events like SVB shutting down stunt the AI growth within the U.S. and allow China or the EU to dominate the space? In a recent research study performed by datainnovation.org, they ranked the U.S., China, and the EU on a few key areas of AI development: Talent, Research, Development, Adoption, Data, and Hardware. The U.S. was ranked first in all categories except Data and Adoption. China ranked first in those. To draw back on my last paragraph, if data truly is king, and China has the data, will they win the race?
Conclusion
So what does winning the race mean? Considering our limited knowledge of the true implications of AI, I think it is almost impossible to tell. In terms of specific development, I’d point to the myriad of blog posts that others have posted that specify movements going on in specific industries. I think it will be a constant power pull between governments, however. For example, I believe AI in cyber-security will be a crucial area to win. Might have to make a separate post on this. Whoever is first in this space will be able to infiltrate governments, big tech, consumer data, media and information platforms, and much more. At the end of the day, similar to what Joseph mentioned in his blog post, I think we are lucky to learn about this space while it is so young. I also think of it similar to March Madness, however, I think we just started the first game of the round of 64 and there is much more madness to come.