From the course: Artificial Intelligence for Marketing

A word of caution

From the course: Artificial Intelligence for Marketing

Start my 1-month free trial Buy for my team

A word of caution

“

- Artificial intelligence is tremendously useful for marketing, and I'm glad you're interested. Rest assured that AI is not going to take over the world and enslave mankind, it's not going to take your job. That's just not going to happen. It is going to take over some of your tasks so that you can do your job better, faster, and cheaper. But there's some things that you need to be aware of when you want to use AI. I can help you avoid some of the speed bumps and the potholes in just a couple of minutes. Start by remembering what economist Leo Cherne said. "Computers are incredibly fast, fast, accurate, and stupid, "while humans are slow, error-prone, and brilliant. "The marriage of the two is a force beyond calculation." Yes, even artificial intelligence is dumb. It can only do what you tell it and it does not know if the answer is reasonable. It's really good at telling good from bad, but it hasn't got a clue about right from wrong. If you tell the machine that you want more email opens, the algorithm might figure out that sending everybody 1,000 emails gets more of them open. That could be true, but it's not helpful, it's not useful. So you have to be very clear about your goals with artificial intelligence. What specifically do you want it to do for you? Next, you've got to give the machine the best data you can. Good data is clean and reliable and abundant and of the proper variety. You may have a million customers, but it's the variety of what you know about those customers that matters. Name and address won't tell you very much, but name, address, time of day they contact you, and what they look at on your website starts to get predictive of what and when they might buy. Too little information is not helpful. The machine will just be highly confident, but wrong. Too many data types, and the machine won't be confident at all, it's just all noise. It's up to you to find the right assortment of data for the machine to answer your questions well. The other troubling thing about data is that it is biased. A group of data scientists at Amazon asked the machine to filter out the resumes of people who were not likely to be successful Amazon employees. Success was I find based on a lot of factors, including the number of promotions, the number of direct reports, glowing reviews, years of service, the usual. But the results were a pile of resumes of only white males. The bias built into the data uncovered their diversity problems. So Amazon had to throw out that algorithm to guard against future bias. Finally, this is a tool, and all tools can be used for good or ill. Should we be afraid of AI? Not at all. Should we be we afraid of bad actors using AI to defraud people, mess with our election process, turn a blind eye to discrimination? Yes, we should. So it's on you to make sure that your projects are well-defined, your data is diverse, and that you are making sure the machine doesn't do the wrong thing by accident or fall into the wrong hands. Keep those things top of mind and you can use AI to become more valuable to your company and make your company more successful.

Contents