Some people use intentionally “wrong” answers as their key to indicate that the call is legitimate. Don’t use questions that could already be compromised or guessed. Maybe arrange for a specific wrong answer – like the car they wanted but didn’t have. Today, some patient portal software platforms include AI answers to questions patients can submit. Today, a few weeks later, the answer is still wrong, but wrong differently. To summarize, ChatGPT says that its answer was based on earlier restrictions – but that’s NOT true either because this interaction was the week before. When challenged, it replied that it had based its answer on earlier conversations, NOT on reading the article – so essentially ChatGPT fibbed to me. Sometimes, it’s difficult to know when your AI assistant is accessing websites and when it isn’t – because AI tools are inconsistent. The tools literally get better every day, and a week or month in AI is like a year anywhere else. This is wrong and is a compilation of various flavors of erroneous information from the web. I don’t have to explain the pitfalls inherent in that approach. I think that’s the benefit of combining the search engine that has years of ranking experience with AI. Google even added a few local features that I hadn’t requested, which I found helpful. I know that Google can easily show multiple locations and stops on maps, but this time, it didn’t show “the map below” that it said was there. I already know where Gyles is buried, and I’ll get to that when I’m not focused on his land in Brewster. When https://www.dnaxplore.com/ ask for a proofreading sweep just prior to uploading to WordPress, I don’t want my content rewritten. ChatGPT has remembered our interactions and, to some extent, has learned what I want based on both instructions and my repeated use of the tool. By now, you might think I don’t like AI, but that’s not true. The prompt is short, only two sentences, but the AI tool generated nearly 50 lines of HTML, CSS, and JavaScript. The AI tool, generates code on your behalf based on your human-language description of what you want it to do. I never considered that I might find cousins through blogging but that’s worked marvelously – both when I publish the articles and later too. I look forward to in-person conferences again, but don’t worry – I’ll continue researching, writing, and covering topics in this amazing industry. First and foremost, I don’t have any intention of stopping. I was explaining DNA topics so often that I figured if I wrote the answer once as an article with pictures and graphics, I could save myself (and lots of other people) a great deal of effort. They will also be able to learn something about the time in which we live today. If you have any questions, comments, or recommendations concerning the archiving of your website, please email the Library’s Web Archiving Team at In the crazy days leading up to RootsTech, a gem of an email landed in my inbox. Plus, I don’t want my articles behind a paywall anyplace, especially since I’ve made them available for free. Resources like FamilySearch aren’t meant for publications like my blog, and neither is WikiTree, especially “someday” after the blog link is no longer valid. Furthermore, “forever” in technology days could be very short indeed. The British decided to secure the three bridges, including the Old North Bridge. From the top of the hill, the entire terrain for miles was visible, and on a clear day, one could see all the way to Boston. As dawn broke, Barrett assembled the armed militia in the town center, which assuredly included Nehemiah Hunt, who lived in the family home that still stands on Punkatasset Hill today. Everything the British were seeking was hidden on the farm of Col. Barrett, who was the head of the local militia units. Worse yet, you really can’t tell and may have no way of knowing whether the communication is legitimate, so you go ahead and click because you’re tired and “it’s probably fine,” except it’s not. AI makes everything easier – including crime. The threat landscape has not improved for us, that’s for sure. That’s another avenue for information gathering, including your employer. Don’t expose any more of your personal information than is already out there – and you might be surprised just how much there actually is. Just when you think you know what to look for and how to interact, you discover that you’re wrong. It doesn’t understand ethics or possess concepts of right and wrong. It gave me the correct answer, indicating that it searched the web and provided a link to a reporting form. This pricing approach is common across AI services today, although every vendor is somewhat different, and the landscape is evolving rapidly. This incident was covered by local news and wasn’t a one-off “sad story” social media plea that couldn’t be verified. You don’t want to share and play a part in victimizing your friends. The really sad part is that truly legitimate organizations suffer because of the level of distrust we must maintain today. Here’s what Google’s AI-generated search summary has to say, along with links to the sources it provided. Yes, I know it’s a pain, but not nearly as big a pain as the alternative. If you have already clicked on a malicious link as described under Rule 1 and entered a password, change that password immediately. There’s even more information available on the dark web, including passwords from data breaches and the names of family members. Whether you know it or not, your email address and phone number are already circulating publicly and are easily accessible. Clicking DOESN’T DO ANYTHING FOR YOU, but it does provide information about you to other people, including AI bots and the social media platform itself. Another effect of reading one of those scammy but very interesting and engaging stories is that Facebook now uses AI (how ironic, right) to show you more of the same. The key, however, is that whatever you watch or read, you’ll see more of it in your social media feed. The phenomenon is often referred to as a social media bubble or echo chamber because it reinforces what you believe based on what you’re seeing and continue to see. I selected this version because it’s actually what I was trying to say and couldn’t find the right words. ChatGPT suggested 7 replacement words, plus three rewritten sentences, including the following. Given that ChatGPT is a LLM, I often ask it language-related questions. ChatGPT can’t always “see” or “understand” images, placement, bolding, or colored text in documents, so it’s easy for it to miss nuances that are covered there. For centuries, the Scythians ruled the vast Eurasian steppe from the Black Sea to the Altai Mountains. The origins of the early Slavs remain one of the most debated questions in population genetics. An Alsatian shares more common ancestors wit... The Roman Empire, at its height, ruled over roughly 60 million people across three continents. Wonderful to host Dr. Jyothis George today - a dear friend, exceptional leader, and inspiring colleague. I don’t want it to change the meaning of or rewrite anything. I mostly use it for proofreading and editing where I’ve given it very specific instructions. In less than a minute, I had an answer – around November 2, 2027.