Insights, commentary and news

Artificial intelligence, or AI, has been around longer than we may realize. I support the argument that the introduction of the first handheld commercial calculator could be considered the introduction of AI. It was revolutionary in business, allowing for complex mathematical formulas to be calculated without paper and pen – not just by early computer scientists with computers that took up entire rooms.

Simply put, today’s consumer-facing AI tools provide human-like outputs based on a prompt of some sort. This prompt could be a description of an image we want generated, or an idea for a written text piece. I took a look at today’s publicly available AI tools from an agency marketer’s perspective.

OpenAI is behind many of the tools available that power marketing platforms such as ChatGPT, DALL-E, Copy.ai, ContentBot and Synthesia. Stability AI powers another popular image generation tool called Stable Diffusion. Various behind-the-scenes AI models also power the accuracy and predictive functions of programmatic digital advertising that result in the sometimes-scarily-accurate targeted ads you may receive.

Copy.ai is a platform that uses artificial intelligence to assist in writing various forms of copy, and it is very much geared toward marketing and advertising. There are currently nine tools within the Copy.ai platform to help write copy for Facebook, Google and LinkedIn. To test it out, I added the metadata from Marcus Thomas’ website into the copy generator for a fictional piece of Facebook ad copy for our agency, Marcus Thomas LLC. The ad copy I was provided was interesting and – spoiler alert – none of it is factual:

“The agency which is led by CEO, Marcus Thomas, has worked with over 100 clients across a variety of industries since its beginnings in 1999. They have recently been recognized as an Automotive News Top 100 Workplace and as one of the 2017 Best Places to Work by Crain’s Cleveland Business.”

OpenAI’s DALL-E tool generates images using word prompts or can create variations of an uploaded image. It was recently made publicly available and can replicate various artistic styles. My own initial testing of the tool left me nearly speechless, while other images I generated fell flat with impossible and inaccurate visual gibberish. DALL-E was used to generate the image for this article, using the prompt “A middle-aged woman writes an article on her computer about AI while working from home in pajamas while drinking coffee.” I’ll let you make your own judgment about the image output of this prompt.

Topaz Labs has several image tools that can enhance photos using artificial intelligence. A low-resolution, grainy photo or video can be enhanced to a sharp, high-resolution file using their tools.

Synthesia is a tool that sits somewhere between content creation and AI. The platform allows users to create videos from text scripts that can be used for tutorials, ads or other videos where a human reading a script is needed. The platform currently has the capability to produce videos in more than 60 languages.

ChatGPT is a recently released tool that is currently still in the research phase. It provides conversational text responses in various tones and lengths. I used ChatGPT to help me generate some email responses and was highly impressed with this tool. Currently, Microsoft is rumored to be investing heavily in ChatGPT. While that could open up incredible integrations with our everyday, Microsoft-centric workflows, my top hope is that Outlook can integrate ChatGPT as a native tool sometime soon to help me reply to my growing list of flagged emails (with careful human review before hitting "send" of course.)

Read the full post, including a review of threats and some experiments conducted by the author.