Golin’s 2024 AI outlook

Wrote a prediction for the agency’s yearly predictions piece. You can check out all the predictions here.

The agency’s presentation isn’t available as text, so for those of you who might be using screen readers or prefer to read, my piece replicated below:

Applied models for all. 2023 was defined by everyone who didn’t already love chatbots learning to love chatbots. Thanks to OpenAI’s ChatGPT, Google Bard and Microsoft Copilot, chat became the one user experience touchpoint to rule them all, right?

Wrong! While those interfaces were powerful in helping new users and folks who aren’t building AI understand what large language models (LLMs) can do in addressing their day-to-day needs, curiosity, or just make conversation, 2024 is the year these models leave the browser and the chat interface and start weaving themselves into every interaction on your device.

And it won’t just be one model, either. From next-generation smart assistant to new media capture and autocorrect—transformer-based models and LLMs are just going to be another layer of 2024’s operating systems.

Suddenly your devices feel more personalized, more context-aware, more powerful than before thanks to a. balance of both LLMs and smaller, more specialized models running on-device.

For PR and marketing professionals, this is going to mean a lot more interaction with AI—even if there’s not a chat box to type into. While less splashy than a new chat interface, the efficiencies here are no different. Look for content tools that Strat nudging or generating alt text for images, for smarter corrections and tweaks based on internal style guides, even context-aware recommendations for work streams that previously required collating across several tools like vetting and influencer management.

The kicker? As these models become more focused, more context-appropriate, they also become more secure for enterprise and professional use. PR pros should have already been aiming for encrypted, approved tools that don’t donate client data to model training, but embedding these brings additional security that transforms how we live, create, and work.

That shift not only means more folks using AI day-to-day, but also using it securely at a time when companies like OpenAI are rolling out legal indemnification. While 2024 might not be the year every company gets cleared to do more with AI, it’s safe to say the tooling making it possible will continue to roll out.

You’ve seen the demo of Copilot writing an email for you, now imagine if Copilot understood you the way your smartphone did. 2022 and 2023’s remarkable pace of progress will look quaint faster than we think.