People Becoming More Thoughtful About AI
No, I don’t mean that people are going to go back to living in caves, starting fires to keep warm by striking rocks together for sparks and killing and eating wooly mamouths.
But what we are seeing is that as people start really digging into AI, there are a variety of security and operational issues that are important.
Number one may be where is my data and how is it being used. This question is probably a bit late, but, I guess, better late than never.
For countries, the term is “sovereign” AI, meaning they, in their case, as a country, are in control of their data. But for you as a corporate user, the issue still applies.
For example, if your users are using free, consumer AI platforms, and they likely are, and then they are uploading sensitive corporate and customer data, you have now lost control and may well be in violation of the law and violating the terms of customer contracts. This is your version of the meaning of Sovereign AI.
For countries, and probably for you, it is about making sure that you have sufficient leverage in your use of AI that you have a chance at controlling things.
At the London tech conference they are talking about the fact, for example, that most likely the data center where there AI is processed is in a foreign county.
It means that an operational AI may now be performing tasks without asking you.
Likely, you can’t do it alone, but it is important to understand what control you are giving up and making sure it fits into your risk tolerance.
Another important factor is infrastructure. As AI moves from prototype to production, if you are running AI internally your will quickly see that you are running out of IT “steam”. If you are running in the cloud, you may not like the bills you are going to see and if you are using third party services like Claude, ChatGPT or Salesforce, you are likely going to see that the all you can eat model is going away.
If you are not sure how to tackle that beast, please contact us.
Credit: Computing
by 