Voice of the DBA

Tag: AI

  • The Evolution of AI

    I saw a study recently where an AI system was used to analyze code and trying to decide if authors were good. The conclusions were things we’d expect, and quite a few people laughed about this on Twitter. After all, if AI comes to the obvious answer, is it useful? Perhaps, but it’s also a little disappointing.

    I’ve felt that way about AI for some time. Years ago I went through an AI demo for the Titantic data set, coming to the conclusion that the lower you were in the ship, or poorer, the more likely you were to have died. I went through a tutorial with Microsoft on flight data, trying to determine why flights were late. It turned out the later they left, the more likely they were to be late arriving.

    While this might seem silly, and perhaps obvious, I think that’s OK. We don’t expect these early AI systems to be better than humans, at least I don’t. However, what I do expect is that they can do at least as well as humans, and having them return results that we could figure out is a good step. To me, this also gives me confidence that a computer can be used to analyze a complex problem or situation.

    There is more evolution coming in the AI and ML worlds. I have no doubt that these systems will improve, and likely find niches where they are valuable and very helpful in making decisions. The world is gathering more and more data, and it’s becoming hard for humans to analyze it, and maybe more importantly, react quickly. The rate at which data can change is overwhelming, especially for humans. I think that we will need computing systems to help process the vast amounts of data we continue to accrue.

    While I don’t know that I want AI/ML systems making decisions for me, I do think they can help reduce the burden of looking over data and help humans to better focus their analysis in certain areas. A symbiosis of computing software using AI, data, and a human to closely watch it is what I think will really help us become even more efficient and focused in our particular organization.

    Steve Jones

    Listen to the podcast at Libsyn, Stitcher or iTunes.

  • The AI Manager

    One of the advantages of a computer is that it will handle repetitive tasks very well. That’s one reason the DevOps world pushes for automation of simple tasks, like compiling code or copying files between machines. We know the software will perform the task the same way every time, giving us a reliable, repeatable process. Even in many AI systems, the structure of the program ensures some level of reliability, though the actual actions or results may vary dramatically based on the inputs.

    For some tasks, this is great. If we have a system watching for a change in code to rebuild the application, that’s certainly a job for a computer rather than a human. If a sensor needs to be checked and an action taken if it exceeds a value, certainly the watching ought to be done by software, though the actions may or may not be something we want a system to do without human input.

    In some sense, I likewise think that the use of AI to manage workers is a bad idea. This was something I read about in Manna, which seemed exxagerated when I first perused the story. Now it seems that some of this is coming true in this story about workers in a call center being manager by an AI system. There is a software on their machines that tracks their activity, the way they interact with customers, and more. If there are potential problems, the software tries to give them messages about how to alter their behavior. Many employees aren’t bothered by this, but I wonder if that will be the case over time, especially if this software starts to affect employment and pay decisions.

    I think this is micromanagement at a level I’d never want to work under. I know that some of these jobs are akin to factory jobs that make widgets, with employees doing the same thing over and over again. However, there also is some creativity and thought required, otherwise I’d assume an AI could just do the whole job. That might be coming, and I wouldn’t be surprised to see it, but as long as humans are working here, there ought to be plenty of human oversight.

    Could you see this in our industry? Maybe AI reminders that you’re slow to close a ticket, or not committing enough, or even that you’re not doing enough testing. It’s crazy to think that there could be software that companies use to manage developers and other technology staff.  My guess is most of you think this would be a poor idea, but I could see this becoming used in help desks and infrastructure at some point. If it works there, who knows what types of AI software might be deployed by companies.

    Perhaps more worrisome for many is that if this becomes a trend, this might affect the middle management levels in companies more than workers. One manager might do the job of 10s,  or even 100s, which is its own societal implications if we use less people as supervisors. Maybe this would make for more efficient companies, and hopefully, less meetings. I don’t know if I think that’s a good evolution of software being used to manage a process.

    Steve Jones

    Listen to the podcast at Libsyn, Stitcher or iTunes.

  • Giving Computers Ethics

    I was reading a fascinating paper recently about autonomous cars. I’m actually excited about having a car that can drive itself, though I think this is likely quite a few years away, despite the hype. Ever since I read Red Thunder, I’ve thought that we would first get full time autonomous cars that would either be limited in where they were in use, or part-time autonomous cars that could only be self-driving in certain places. Dense inner cities, or maybe isolated highways might be good places to try this, in my mind.

    While we want to do some programming of these cars, we also have a lot of AI/ML systems in place that run models trained to react in certain ways. They identify things that are moving and stationary, trying to determine how the car should navigate and react. The systems aren’t quite as tightly programmed as many of us expect, with if this then that logic. Instead they have guidelines that are decided upon by the designers and then reactions to data inputs and analysis are a little more fuzzy.

    What are the goals? Well, in most cases they are just moving the car safely down a road. In crisis situations, it’s a little more murky. What happens when collisions are unavoidable? How should the car react? Humans often panic and do strange things, but we don’t want erratic behavior from automated systems, so what should we set as goals? There’s a bit of research that was done to ask humans what they would do when they can consider the situation a little more slowly.

    In short, humans make different decisions in different cultures. There are clusters and tendencies in different parts of the world, which is interesting. While people are people and behave similarly in many cases, we tend to value different things, depending on our views of the world. That can be problematic when we start to expect computer systems to be more consistent or predictable. After all, we should decide how computers react and be able to trust our decisions are followed. It is up to humans to imprint our ethical desires as a society on computer systems.

    This is an area where I feel AI and ML systems are moving faster than our ability to comprehend the implications. I would want to have a framework built for automated systems, certainly cars, and then expect all vendors of systems would implement that framework in their vehicles. However, this goes beyond cars, and in any places where we are using software, AI/ML based or not, we ought to publish a comprehensive outline of the way in which our system works.

    Computers have the capability to improve our world and reduce chaos, but only if we agree on the way in which these systems work, and disclose in a transparent way what data they handle and what decisions they make based on that data. I hope that we start to get better about informing the world the goals and operation of our systems. I’m not sure that will happen anytime soon.

    Steve Jones

    The Voice of the DBA Podcast

    Listen to the MP3 Audio ( 5.6MB) podcast or subscribe to the feed at iTunes and Libsyn.

  • The Creepiness of AI

    Last year, I watched a keynote talk from Matthew Renze about AI. In his talk, there were examples of the amazing things that Artificial Intelligence can do, as well as some of the creepier things have have been developed. It was an interesting talk, one that gives me inspiration and hope for the benefits of better computer algorithms as well as the concerns for various issues that we may be unprepared to deal with as a society.

    One of the more controversial items that occurred recently with AI was the Google phone call, where a computer answers a call and interacts with a human. What’s disconcerting here is that the person doesn’t know this is a computer, and there are speech patterns the computer uses, like um interspersed in the answers, that deceive someone. While this certainly might be helpful in scheduling situations shown in the call, there is a downside. Could you imagine artificial personas used in telephone scams or phishing situations? A help desk knowing some information and then asking for verification of other data?

    There are perhaps greater concerns, such as the work done with imitations of President Obama. There are fake speeches, generated by computer. While movie studios might want fake actors used to reduce labor costs, do we worry about the implications of a computer actually being able to imitate one of us in a video call?

    The use of AI and ML, with lots of data an organization might have gathered could be good and bad, but certainly opens the world to more problems than benefits if there isn’t mandatory disclosure of the cases where this is used. Since there are always going to be criminal elements that don’t obey rules, this might be very scary.

    There are certainly other issues, such as Target predicting a pregnancy, which was the first really, creepy data analysis thing I saw. That one is a few years old, and still bothers me as it was accurate, but an unrefined use of the data. A good example of where marketing groups are a bit too excited to use AI/ML technologies and don’t think through the implications. Fortunately this case seems to have dampened some of enthusiasm for prediction in retailing.

    Perhaps this item that is a bit funny, but it is also very worrisome for me. It’s the case of an AI system playing video games. The AI system decided the best way to get the best score was to pause the game. Rather than compete and try to do better, the computer decided to just stop. A completely unexpected outcome, probably because the feedback and expectations weren’t explicit. Since it seems quite often humans don’t specify their requirements or expectations very well, I could imagine this being a very large issue in AI systems as they are used more often. It could even be deadly or problematic when a system does something we didn’t anticipate, and impacts human health.

    Most of us won’t work with AI much as a technician, other than providing or managing some of the data. I do expect AI and ML systems to touch more and more of our lives, perhaps using our data for good, perhaps not. Hopefully we can help steer applications into the former more than the latter situation.

    Steve Jones

    The Voice of the DBA Podcast

    Listen to the MP3 Audio ( 4.7MB) podcast or subscribe to the feed at iTunes and Libsyn.