AI round-up: Week of Jan. 08, 2024
Deepfake Undermines Premium Brand Claims
TopTier, one of the leading manufacturers of high-end consumer goods, was thrown into disarray this week by a deepfake video making it appear as if popular YouTube product reviewer UnboxDan claims TopTier electronics fail to meet the premium quality standards and performance the brand promises.
In the slick but completely fabricated video, a fake UnboxDan conducts side-by-side tests showing TopTier televisions and sound systems underperforming budget models from discount competitors. The faked reviewer claims he could find no meaningful differences justifying TopTier's far higher pricing.
The deepfaked video quickly went viral, and #TopTierExposed began trending on social media as real-world UnboxDan clarified he never created such a review. But the damage was already done. TopTier stock slid 9% amid consumer outrage before the company could officially respond. Retail partners demanded an explanation for the alleged inferior quality exposed in the video.
While denying the veracity of the claims, TopTier admitted its investigation into the Deepmaster-fingerprinted video's origins suggests it was created by a recently fired R&D engineer.
With a brand reputation on the line, TopTier is working urgently to contain the fallout from the synthetic video. But the crisis revealed one disgruntled insider could unleash major turmoil by weaponizing TopTier's premium brand claims against itself. The company now faces tough scrutiny around the transparency of its performance marketing and why its risk analysis overlooked the glaring vulnerabilities posed by deepfake technology.
So, if you haven’t suspected it already, this is a fake news story about something terrible that could happen. (Note: I used a prompt to create this fake scenario within Claude.) It could happen sooner than later. It could happen before we talk next week. I’m working quickly to update client crisis offerings to include planning for scenarios just like this. Especially after seeing that Ethan Mollick did this late last week.
But wait, that’s not all the news we have this week…
This week’s issue: 9 stories, one fake story (above), one must-see LinkedIn video from Ethan Mollick (above) and three podcast links
The AI impact meter: High
How long to read: I'm not even gonna guess this week …. it depends on how much you dig into because there’s a lot.
1. Actors union announces deal for AI voice acting licensing in video games. Remember when we talked about the strike ending, and I said this was next? Because why wouldn’t it be? If you’re an actor, you can now live forever. Be in multiple places at once. And make sure you’re not replaced by AI…by becoming AI.
2. Google is working on an advanced version of Bard that you have to pay for.
3. There is significant concern about fraud, deepfakes, and misinformation using AI in the large number of elections worldwide this year (66 in all). So, this NPR article highlighting how election officials in the U.S. are preparing/planning is very interesting.
4. CES is happening as we speak, so I’m sure I’ll have some round-ups to share. However, one preview item caught my eye – an AI-driven wellness mirror that analyzes your face to see how you’re feeling.
5. The OpenAI Chatbot store is back on. Stay tuned. It’s supposed to open next week.
6. Speaking of OpenAI … they have released ChatGPT Team. You can now get your whole team/organization on it for a licensed agreement, per user/per month.
7. Pennsylvania government workers will start using ChatGPT in test program.
8. Let’s talk about the e/acc movement for a minute.
- So I’ve heard the term a few times on The Hard Fork podcast. And you may have seen “e/acc” appearing on social handles lately. It’s a movement and belief that stands for effective accelerationism.
- The definition: Effective accelerationism embraces technological progress and the dynamics of capitalism as catalysts for change rather than advocating for a reduction in economic growth.
- Last week, I heard it mentioned (as you may have since we called it out in the last AI Round-up) on the Lex Fridman podcast when he was speaking with Beff Jezos, aka Guillaume Verdon, who started this movement. He spoke at length about it.
- This week, on the always resourceful Marketing AI Show, Paul and Mike discussed that pod and the e/acc movement in greater depth. I would highly recommend listening to their breakdown and explanation of it.
- Just for reference, here is an article from The Information, back in October 2023, that talks about it … where people featured in the story go as far as to call it a cult.
That’s all for this week. I have some stories that got bumped to next week, and I’m sure we’ll have some CES wrap-ups. Note: I’m always available to talk or dig deeper offline on any of this. That’s why I’m doing this – give me a call or an email, and we’ll talk!
As a reminder, this is a round-up of the biggest stories, often hitting multiple newsletters I receive/review. The sources are many … which I’m happy to read on your behalf. Let me know if there’s one you’d like me to track or have questions about a topic you’re not seeing here.