Modern Life

Artificial intelligence comes to the office

The latest technology should excite workers, not concern them — if company leaders learn to let their employees safely play.

By Heidi Mitchell — June 29, 2023

Until a couple of months ago, creating marketing messages for global surrogacy agency ConceiveAbilities was madness. Its staff would spend hours, even days, individually crafting and refining emails, blog posts, and other communiques, taking the organization’s main message and repositioning it to four distinct audiences: target egg donors, surrogates, intended parents seeking an egg donor, and those looking to match with a surrogate, says CEO Cathy Kenworthy. 

Then came generative AI: Now the marketing team just asks it to rewrite the statement for each subgroup, using topical contexts. Kenworthy loves incorporating generative AI into her organization’s everyday tasks, since it saves her marketing staff hours of time “so we are able to focus on pressing initiatives.”


RELATED: What is… artificial intelligence?


Meanwhile, Taj Reid, global chief experience officer at Edelman, a global communications firm, says that many companies are integrating generative AI into their daily work. His firm recently partnered with a mayonnaise brand to create a campaign driven by AI. Customers could type in what was lurking in their refrigerator, and the AI would generate recipes that aimed to reduce food waste. To emphasize the human element, Edelman partnered with chef and non-chef TikTokers to lead live cooking demos of those very dishes. “That was a fun application of what AI can do,” he enthuses.

He’s also seeing companies use AI to summarize large volumes of data and creatives turning to AI to eliminate that dreaded blinking cursor. “If you’re freestyling, AI can really provide context. If you wanted to pitch a story idea to the New Yorker, for instance, you can tell the AI to pretend it is a writer for the magazine, whose audience is a busy executive. Then you can ask it to take a passage you wrote and make it more eloquent, in the New Yorker style... and it returns a polished pitch.” And, he adds, AI capabilities and use-cases are expanding every day. “Today is the worst AI will ever be,” he says. 

“I don’t believe that AI will take your job. I believe that people who know how to use AI will take your job.”

— Taj Reid, global chief experience officer, Edelman

It seems like every day there’s a new generative artificial intelligence, from Bard to Bing to another iteration of OpenAI’s ChatGPT, coming to market — along with countless, breathless headlines hyping its benefits and infinite possibilities. And while asking an AI trained on a large language model (like ChatGPT’s use of the entire internet, through May) to compose a poem for Father’s Day may feel magical, less media real estate has been devoted to workplace use-cases of this emergent technology. 

Sure, generative AI can be glitchy and certainly requires some company-wide guidelines, but this creative tool est arrivée, and thought leaders need to be considering how to employ it for the enterprise — whether that means generating meeting agendas, turning notes into presentations, creating sales or marketing assets, framing a pitch, or outlining a business plan in seconds. They also need to recognize its potential pitfalls and set some ground rules.

Old technology, new use cases

“AI is not new, it’s been around for decades,” says Scott Hallworth, HP’s chief data and analytics officer. “Chatbots and Siri and Alexa and Grammarly came about a long time ago. What’s different now is the speed and scale, and the ability to optimize the training model.” Hallworth spends his days thinking about AI and has found a few talents at which technology excels. 

Firstly, he says, generative AI — which essentially recognizes patterns — is very good at search, especially within a locked company ecosystem. “Documents, code, libraries, whatever it may be, you can now search for that fast and with precision,” Hallworth explains. Content creation is the second-best skill of a generative AI. “It’s very, very good at translation, for example, which once took hundreds of hours for an international company like HP, especially when trying to get the colloquialisms right.” Coding, too, can easily be generated by an AI as a starting point, since code is typically black and white, right or wrong, says Hallworth. 

While companies are finding massive time-saving use-cases for AI, individuals in advanced roles, too, are finding ways to enhance their own skillset, he says. “It’s an augmentation and an accelerant… you still have to edit and review its output, but the speed in which I can get something done is a lot faster than if I’m sitting there staring at a blank sheet of paper.”

Fabio Buonocore

Pitfalls on parade

Scores of journalists have reported on the “hallucinations” that an AI can suffer, which is a consequence of the tech having such a talent for guessing. Less discussed are privacy issues. While Hallworth and Reid are big advocates for using generative AI to speed up and simplify work tasks, both believe companies need to set some rules of engagement before letting employees run amok with the tool. “I wouldn’t put any sensitive information into an AI like ChatGPT now, because you are offering that up to an open community,” says Reid. It’s not as though the AI companies are actively stealing your intellectual property, he explains, “but there is potentially an engineer who might be looking at inputs of large datasets and come upon your next, secret product launch. So be very careful when you use tools that are open source. My recommendation is not to plug in personal or sensitive intel into an AI until someone in your company has vetted that specific AI and whatever it is you’re inputting.” 

Hallworth’s approach is more structured and safeguarded. He allows employees to request a private “sandbox,” to which each user has a specific key, permission, and access rights. The data is synthetic or a sampling and used to see if an idea has any legs. “The key thing is that the sandbox operates like any other product funnel. You’re going to go from ideation to exploration to development to creating code,” he says. “And if there's any there there, you pass through a series of gates into a production environment to then be monitored and managed with all of the company’s cyber and legal controls.” If they don’t make it past that gate, the sandbox and all its data is permanently deleted after a relatively short period.

In the not-so-distant future

Reid foresees a time when an AI can generate designs that can be 3D printed at scale and deliver goods quickly to those who want or need them. He sees a nearer future when workers can use AI to understand the other side of an argument or ensure their product and practice and messaging adhere to the most recent laws, as ConceiveAbilities' director of content marketing Marci Hughes recently did. 

“I used a generative AI to share our blog about New York State laws, tying it in to Andy Cohen’s recent headline news about his journey with one of the first surrogates in New York state,” she says. “I typed ‘Andy Cohen Surrogacy Laws New York’ into an AI caption writer, then tweaked the post based on our needs.” 

The upside for workplace use-cases is almost limitless, and nothing to be afraid of, insists Reid. “I don’t believe that AI will take your job. I believe that people who know how to use AI will take your job,” he says.

He reminds his team that we’ve been here before: new technology comes along, everyone panics that they’ll soon be irrelevant, and even better jobs are created. The most critical advice he has for leadership is to allow their teams to engage with this new tech, to get to know its best uses. 

“It’s critical that people know about AI and immerse themselves in it at work, though safely,” he says. “I see AI as an accelerant, of both good and bad ideas, but if you don’t participate, you will miss out. It is very much of the now.” 


What are… robotic process automation and chatbots?