TJ Walker (speaker) warns that synthetic audio/video is outpacing truth, giving examples where AI produced erroneous or fabricated content (Washington Post’s personalized AI podcasts) and where politicians have weaponized AI clips (White House example, Gavin Newsom’s AI videos). He confesses to using AI for drafting copy and maintaining an AI voice/video clone that can err, and he argues this trend threatens democratic processes—especially in close elections—so higher-level regulation and standards are likely needed. The piece mixes reporting, personal anecdote, and a call for audience feedback.
Questions this video answers
Within about 48 hours of posting.
Chapters
When AI talks politics, who can you trust? The Washington Post has rolled out personalized AI generated podcasts, but the launch has been rocky. Semaphore reports that the bots have mispronounced names. Now, confession. I've done that occasionally, so I guess if human beings can do it, AI bots can do it as well. They've misattributed quotes and even invented lines, prompting alarm inside the newsroom. The Washington Post head of standards called the errors quote frustrating for all of us, while one editor wrote that the product deliberately warps its own journalism at scale. At the same time, AI is fueling sharper partisan warfare. After the White House used SZA's big boys over an ice cuffing session video, California Governor Gavin Newsom answered with his own AI generated clip show showing Donald Trump, Pete Hegseth, and Stephen Miller in handcuffs. You've got to look at. I'm not going to show it here for a variety of reasons, but look at it. It is disturbing at some level. Some of you may find it wildly amusing, some of you may find it troubling, some of you may find it both. But it is a sign of time to come together. The episodes show newsrooms and politicians racing to weaponize synthetic audio and video, while accuracy, consent and trust struggle to keep up. Okay, so this relates to one of our stories earlier today about Disney giving away all of its characters to OpenAI. We are seeing company after company, big, medium and small. I guess I'm included, giving their content to the AI powers that be. Creating content, hoping to cash in or at least not become completely irrelevant. Now, the Washington Post has a lot of very articulate, intelligent people who do podcasts. Washington Post even has its own fast TV network. So they're used to creating video and audio. They have their own podcast, but they want more. And I have to think, while I've not talked to their publisher, they're thinking, we got to prepare for a day when all of a sudden other people are putting out 10 million AI podcasts a second. We're going to get lost. We're going to drown in that. See if we're not also creating a lot. The problem is sometimes the content messes up. Now, confession. I know this firsthand. I have my own AI generated video audio clone. Sometimes it gives information that's not correct. Now I correct it. If I'm alerted to that, I can go in and say, don't ever say that. I can correct it, but sometimes it's made mistakes. I apologize. I let you. I do let you know it's a possibility. Now, what you're watching right now and what you're listening to right now is coming from a real person. I'm TJ Walker. This is real. We're doing it live on TikTok, YouTube, Facebook, every other social media platform. But confession time. I do use artificial intelligence to write, typically the first 120 words of a segment of a news segment. I am feeding it a story from a reputable news source like the Wall Street Journal OR Reuters, asking ChatGPT to rewrite it, condense it, give attribution to the source, and then I use that to summarize the story to start off the conversation. And I read typically from a teleprompter. I don't have one today because I'm on the road, but it is occasionally messed things up. It is referred to President Trump as former President Trump. Well, he's not former President Trump at the moment. I realize a lot of people are looking forward to him gaining that title again, but it's factually incorrect and I've caught it before it came out of my mouth. So I try to correct things. But we are living in a time when more and more big companies and individuals are using the power of AI to create content. And the real danger, of course, I mean, the biggest example of danger is a leader of a superpower seen in a video in a way that looks like it's completely new and real, coming from the White House or from the Kremlin, saying, we begin bombing an enemy and everyone thinks it's real. And then that country said, well, time for mutually assured destruction. Their leader is saying, they're bombing us. We have no choice, we have to bomb it. And it was all AI generated. Now, you could say maybe there are intelligence forces within governments that are powerful enough to have nukes to check on that, but it certainly could cause riots, it could cause hysteria, it could cause. It could cause lynch mobs to really attack people. Now, as much as I find Governor Gavin Newsom entertaining, interesting and articulate and deserving the position right now to be seen as if not the leader of the Democratic Party, sort of certainly one of the top tier candidates for 2028, and certainly someone who deserves title of a leader of the Democratic Party, I am a little bit troubled by the use of AI video showing current political leaders in power doing things that they're not doing. Now he's showing President Trump, War Secretary Pete Hegseth, Stephen Miller, handcuffed, sitting in the street, sitting on a curb, being taken away to jail, presumably in the back of a car. And obviously this hasn't happened. I don't think it's going to fool that many people. And depending upon your politics, it's wildly amusing. My concern is for other political campaigns where the candidates are not as well known. Let's say it's a highly contested congressional race and let's face it, in most congressional districts in America, people don't even know anyone running for Congress. They don't even know what they look like. So if it's a close race and at the last minute, let's say I'm running against you and you're a great person, honest, ethical, and we're completely tied in, in the polls and I'm panicking because I don't want to lose and I just show you a video of you beating and lighting your next door neighbor's dog on fire. And I just use a video to do it. And I'm sending this out the night before the election. Look what the, you know, candidate Smithers has done. Vote for tj he loves dogs and he'll take care of your dogs. That could change votes. People would be so outraged again if they don't know, if they've never heard of you, they don't know me, they don't know you. And they see this video of you and it looks just like you, a real person. And I perhaps put a nice news label. So it seems like the most reputable news organization in your town. Showing it, it could change a lot of votes. It could change mind and in close elections, it. You don't have to change that many votes to change the entire outcome. So this is why I think we may need some regulation. Regulation doesn't solve every problem. It cannot get rid of every piece of unethical activity in the world. But this gets down to fundamental pipes, if you want to think of it that way, of democracy. If people don't really know what's going on with candidates with their positions, their statements, their facts, their accomplishments, their, their bag. If they don't really know what's real, then the entire premise of democracy and voters making informed decisions is out the window. That's the real problem. Let's look at a few more facts here going on with this, with this issue. Regulatory and industry standards for AI use in journalism and politics remain thin. There's no doubt about it. It's going to have to, there's going to have to be some regulation, at least at the higher level, because otherwise our fabric of our society could just be ripped apart. Now, we've already seen in other countries. Sometimes it's not even AI stuff. It's just people posting. Hey, people of this religion are killing somebody of our religion and here's where they live. Boom. They are attacked and in some cases, death. It's happened because of situations on Facebook and other social media. The AI generated video could in fact, make it much, much worse. I'll be keeping my eye on it. What do you think you can post your comments? I do typically respond to comments questions within about 48 hours of posting. And I'm a little bit behind this week because I've been doing a lot of international travel. But please, if you hear anything you like, hear anything you don't like, hear anything you want to disagree with, post your comments right now on the show. Wherever you're watching this, whether you're watching on X or LinkedIn or YouTube or Facebook or TikTok or Rumble or any other site, we're open to all sites. I'm here for you.
