What’s News: In the last week, there has been a whirlwind of negative news concerning the big AI players. Google announced it was halting image generation on its Gemini tool so it could fix historically inaccurate images being created — such as an African American George Washington. This news followed a big hiccup at ChatGPT, where a code update resulted in its query results spitting out gibberish. This brings to mind the first rule for using generative AI: doublecheck the results.
These tools are powerful creative forces — with the emphasis on “creative.” Just a couple weeks ago, I was working on a 10-Things-You-Didn’t-Know listicle with a client for their syndicated morning show host. ChatGPT not only said that this DJ had made appearances on CBS primetime dramas, but that he also was a frequent guest-host on Saturday Night Live! Thankfully, these claims were outrageous enough for us to immediately question all of ChatGPT’s claims; but these AI hallucinations are not isolated.
As you continue to include AI in your daily workflow, harness AI’s power for creativity; it can save you time and brainpower and spit out ideas faster than anything. But also ALWAYS doublecheck the results.
I’d love for this AI Edge to be the beginning of the conversation. If you have a tip, product, or an AI hallucination you’d like to share, please reach out to me: chris@g5j.8ac.myftpupload.com.
AI EDGE – TOOLS TO KNOW ABOUT FOR YOUR STATION
Summariz.ing: Paste a YouTube URL and it creates a summary and transcript. Free for up to 60 minutes. Perfect for summarizing school board and city council meetings.
Pika.art: Creates animations from still photos. Premium tier creates lip-synced videos from still images. Good for short social posts and memes.
AI EDGE – INTERVIEW
Jeremy Sinon at Hubbard Radio
Jeremy Sinon has been Hubbard Radio’s go-to-guy for digital for eighteen years. After working in the company’s Minneapolis cluster, he migrated to the corporate ranks and is Hubbard Radio’s VP of digital, overseeing digital operations for the company’s fifty-plus radio stations from coast-to-coast. Sinon sat down and spoke to me about why he is using AI. – Chris
CB: How is Hubbard approaching AI?
JS: We’re trying to formulate organized discussions between business units and departments around AI. The goal is for these discussions to be proactive but cautious, “What’s new? What’s cool, What’s out there?,” and What should we be looking at?” We’re trying to protect overlaps of technology and to connect the conversation together so everybody’s on the same page. Our staff wants answers to “What can we use?” and it’s not a simple answer.
CB: It varies by department.
JS: A salesperson using AI to draft an email versus an editor using AI in a newsroom? Totally different deal.
CB: Many broadcasters are grappling with risk.
JS: There is risk and reward. It used to be if you wanted to use software, you signed up for a license to start using the software on your computer. Then came the cloud, and now the data is not completely just sitting here locally. The next level of this is AI, and it’s not just it’s living in the cloud, but it’s living in the cloud AND it’s training models off of what you create. It’s a lot to figure.
CB: But on the upside are the potential rewards.
JS: We’re in the boat right now of thinking AI can empower employees to do more and create more. Can we produce more content and produce more sales? Can AI help us be more productive? A year from now if we’ve got some things in place that we could measure and say we’ve increased productivity. I think we’ll be pretty happy.
AI EDGE NEWS TO KNOW
New Top 200 Daily Podcast Is Completely AI-Generated
Perplexity.AI, a “conversational search engine” backed by Jeff Bezos is on a tear. Not only is it looking to take a bite out of Google’s search traffic, it has launched a current events podcast, Discover Daily, that is completely AI-generated. The four-minute cast is a pleasant (yet a little dry) version of the news-update you’ll recognize from NPR and BBC. It also has popped into Apple’s Top 200 Podcasts ranker. Definitely worth a listen.
AI EDGE ‘WORST PRACTICES’
Another AI-Generated Courtroom Fail
A Vancouver lawyer, Chong Ke, is facing the heat after citing two fake cases generated by AI tool ChatGPT. These fabricated cases were supposed to help a divorced dad take his kids to China, but they turned out to be complete inventions!
Instead of winning the case, Ms. Ke ended up owing the opposing counsel a hefty sum for the time they wasted trying to track down these non-existent cases. Yikes! The judge, while accepting her apology and acknowledging no intention to deceive, expressed serious concerns about relying on AI without proper verification. Talk about a case study for responsible tech use!
Oh, and the plot thickens! The judge also ordered Ms. Ke to go through her other files and disclose any other instances where she might have used AI-generated materials.
See you next week!
Chris Brunt
Director of AI, Revenue, and Digital
Jacobes Media | jacapps
Along with his duties at Jacobs Media, Chris develops revenue opportunities for jācapps’ mobile apps clients.
- Ten Recent AI Stories You May Have Missed - November 15, 2024
- The AI Patrol is Ramping Up - October 25, 2024
- All-AI Radio Is Here - October 18, 2024
Leave a Reply