AI Top-of-Mind for Dec 12
More end-of-year AI reviews, this time from Nitin Dahad at ‘EE Times’ looking at the good, the bad, and the ugly side of AI. From the article on ugly:
· There’s also potentially an ugly side to AI, as highlighted in a paper to be presented at next week’s conference on neural information processing systems in New Orleans (NeurIPS 2023). It suggests that AI networks and tools are more vulnerable than previously thought to targeted attacks that effectively force AI systems to make bad decisions.
· The researchers from North Carolina State University’s Department of Electrical and Computer Engineering developed a piece of software called QuadAttacK to test the vulnerability of deep neural networks to adversarial attacks. The software determines how data can be manipulated to fool an AI network — and see how the AI responds. Once it identifies a vulnerability, it can quickly make the AI see whatever QuadAttacK wants it to see.
Building on the holiday theme, the continuing shortage of GPUs and the companies that are naughty and nice in Nvidia’s gift book. ‘The Information’ does a good job of listing who received what, and there a few interesting additions and omissions.
Turning to marketing, David Leibowitz in ‘UX Collective’ looks at some of the real-life use cases for Generative AI. From the article:
· E-Commerce: product descriptions
· Engagement: QR codes as digital art
· Experiential Omnichannel Engagement
· Customer Persona Development
· Customer Journey Storyboards
· Visual Merchandising
· Brand Marketing
· The days of static, one-size-fits-all marketing assets could soon be behind us. Instead, we’re moving towards a future where marketing visuals are as dynamic, diverse, and personalized as the customers they’re intended to reach.
· Generative AI is not just a tool — it’s a catalyst for a marketing revolution, and brands that adapt and embrace the capabilities will be at the forefront of this exciting new era.
But there are cautions. Meta is the latest to be developing AI protocols in the hopes of preventing risky AI outputs. Note that they are also one of the leads in the newly formed AI Alliance. You can read why they call it ‘Purple.’
And with all the talk about ChatGPT ‘expert’ prompts, what are the real benefits? Read here. From the article:
“When a user provides a prefix like “You are an expert in social marketing” and then proceeds with a request for help with a viral campaign, I don’t treat the prefix as changing my level of expertise,” clarifies ChatGPT. According to the AI, it serves to set the expectation for the type of advice sought, not the depth of knowledge it can provide in the field of social marketing.
Lastly, ‘Crunchbase’ on the differences between ‘Hard’ and ‘Easy’ AI. Hint… security and transportation are hard. But so are medicine, enterprise SaaS, and even workforce re-engineering.
