Outsmarting AI. Brennan Pursell

Чтение книги онлайн.

СКАЧАТЬ Entries in tables for names, addresses, purchases, etc., are typical of structured data. Unstructured data include texts, posts, images, sound clips, and videos. These files, while digitized like everything else on a computer, are not neatly arranged into rows and columns. An AI system will have to do a lot of calculations to classify what’s in them, but you can train it, and dumb as it is, it works immeasurably faster than you do.

      Without data, AI can do nothing. AI can process structured and unstructured data and present information about it in a manageable way. (You will come back to data in chapter 5.)

      Rule 2: Math Is the Father of AI

      AI is just math! Math and its companion, statistics.

      Coded AI systems are expressions of mathematics and logic. Statistics rely on the same. AI algorithms use calculus and linear algebra to work over data in numeric form to get results. The math can get very complicated and sophisticated, but for all that, it’s still math.

      An honest AI pro tweeted: “It’s AI when you’re trying to raise money, machine learning when you’re trying to hire developers, and statistics when you’re actually doing it.” This says it all. Statistics is just applied mathematics, in AI, for data analysis.

      You may well be wondering, “So if AI is just math, mathematical procedures done on numeric data, then how is AI different from data analytics, predictive analytics, data science, and big data?” Well, they are all part of the same family of algorithms performed by software. A key difference among AI algorithms is their ability to self-optimize—some would say, “to learn.” (We will revisit this important matter in chapter 2.)

      The beauty of recent AI software advances is that you do not need to learn, memorize, and key in the algorithms in order to get the outputs you need from your data. Nor do you need specialized hardware. You can even build your own AI data analytics system online in the cloud by dragging and dropping elements into place.

      Rule 3: AI Systems Are like Kids—They’re All Unique

      For those of you who were offended by the gendered parenting roles in the preceding two sections, please forgive and get over it. We abandon all gender references when it comes to the kids.

      AI systems are like children: Each is unique. You could compare them to fingerprints or snowflakes for the same reason: No two copy each other exactly as they do their work, even if they do the same job and process similar data. And in many cases, we are not quite sure how they actually come up with the results that they do, even if we know the data they come from.

      AI algorithms adjust their inner workings according to the results they are trained to output, given the inputs. In that sense, you could almost call them “organic,” if not really “alive.”

      If you adopt an AI system into your organization to improve one of your business processes, then it will rapidly become your own, unique tool. AI software does not come “out of the box,” and even if you do use a vendor’s software-as-a-service (SaaS), trained on your data, the AI system will really be all your own.

      Rule 4: AI Needs Parenting

      All kids need parenting throughout childhood, from start to finish. Yet unlike children, who we can reasonably expect to grow into self-sufficient adults, AI systems remain needy until the end of their software life cycle.

      In the beginning you will have to do everything. Even if the software is readily available, the data has to be prepared, integrated, and validated. The system has to be thoroughly trained and tested over and over again, and then there is the great work of socialization. All coworkers who come into contact with the system will have to understand it, accept it, and work well with it. Once your system is firmly established, however, you will just need to check in and test it periodically to maintain your governance. If the input data changes in some unexpected way, the whole system could go haywire. If it takes a village to raise a child, it takes an organization for AI to succeed!

      Once fully functional, an AI system can be generally relied upon to complete its carefully scripted tasks, but beware when it comes to decision-making! At what age do you entrust children with authority? Autonomous algorithms left to make decisions can be like little kids with meds and guns.

      Cringe-worthy examples abound. An AI algorithm in Idaho cut Medicare payments to four thousand disabled people, which prompted a major lawsuit. The database it relied on was loaded with gaps and errors. What did the people in charge expect?

      Armed with AI, businesses can make themselves truly destructive if the people give in to natural recklessness. Want a global financial crisis worse than 2008–2009? Just set up AI-powered self-executing credit-default swaps. Want a criminal justice system devoid of reason and humanity? Turn it all over to computers. Want World War III? Turn the US president’s nuclear football into a fully autonomous algorithm.

      Want to see a real, live, AI-powered social media platform make money at any cost, despite all the suicides, extrajudicial killings, child pornography, child brides, and illegal drugs? Take a close look at Facebook, which owns Instagram and WhatsApp.

      Facebook’s AI did not catch the live-streamed massacre in New Zealand on March 15, 2019. It couldn’t. It had either not been trained, or its trained model failed. A Facebook user flagged the gruesome post within minutes, but the company did not react. Only after the police called in, about an hour after the event, did Facebook remove the original video. Facebook, YouTube, Twitter, and Reddit struggled to take down the 1.5 million re-postings of the slaughter.

      People using AI need minds unbent by malice and gross negligence.

      Rule 5: Embrace the Cyborg

      When we say “Embrace the cyborg,” we refer to the institutional and functional level, not the individual. Of course, if you want to implant autonomous control systems into your body, you are free to do so. The biomedical implant industry has been around for quite some time, and recently it has begun to grow in new ways.

      AI can assist life without an implant. Some companies are working on an AI solution that helps the blind to know their surroundings. The blind person carries or wears wide-angle video cameras and other sensors, the system classifies objects captured in the video files and other data, and an automated voice narrates the scene as the person moves through it. It’s wonderful!

      On an institutional level, robots have been assisting people in their work in manufacturing plants and warehouses for years, and their numbers are rapidly increasing because AI makes them much more adaptable to changing circumstances than in the past. Now chatbots are helping as well. AI-powered predictive analytics can help people complete tasks better, more accurately, more effectively, and more efficiently in just about every sector of the economy.

      All throughout organizations, AI can help augment people in their work through image identification, voice-operated controls, automatic data entry and transfer, and other functions. People should not be afraid of these AI augmentation tools, because the software will never be able to replace their critical thinking, sense of judgment, and human awareness and understanding, which in many ways is a business’s most valuable or fundamental asset. People, whether customers or coworkers, are frequently unpredictable, and computers are very bad at processing unforeseen situations.

      Hybrid AI-human systems, combined with expert human policy makers, can deliver real business value for your organization.

СКАЧАТЬ