By Jim Misener, CEO, 50,000feet
OpenAI recently unveiled a significant advancement in the ongoing generative AI arms race: making ChatGPT more human. The newest version, GPT-4o, goes beyond textual responses by recognizing emotions in users’ voices, interpreting their facial expressions, and adjusting its tone and cadence to match user preferences. For instance, if you ask for a bedtime story, it can adopt a soft, soothing whisper.
A more human ChatGPT is not exactly an out-of-the-blue development. There is a growing trend, especially among Gen Z, to turn to AI for personal companionship. Apps such as Pi offer life coaching in a helpful and conversational manner. Replika, one of the most well-known, is a virtual friend and mental health lifeline. Blush is an AI-generated dating simulator. Why is AI becoming more human, and how should brands respond?
Major Advances
Businesses have long sought to make AI interfaces more approachable and natural. Remember in 2019 when Amazon made it possible for anyone to program an Alexa command using the voice of Samuel L. Jackson? It seemed like an amusing novelty at the time, but Amazon was paving the way for voice-activated prompts to become more familiar, and incorporating the distinctive voice and comical attitude of the well-known star was an important step.
Since 2019, the world has changed in some important ways that have paved the way for brands to humanize AI, especially on the front lines of marketing and customer service. First off, there’s the obvious explosion of generative AI and supporting technologies. The availability of greater computational power has allowed for the training of more complex models like GPT-4, which can process and generate human-like responses more efficiently. A proliferation of publicly available digital content, including text, audio, and video data, has provided the raw material needed to train these AI models. Continuous interaction with users provides real-time feedback, allowing AI to learn and adapt quickly to human preferences and nuances.
But there’s something else going on at a societal level. The growing interest in AI companions, driven by social isolation and the need for mental health support, particularly during the COVID-19 pandemic, has accelerated the development of more human-like AI. Users now expect more engaging and natural interactions with technology, pushing companies to develop AI that can meet these expectations. This is especially true of the surging Gen Z population. Recently, Npj Mental Health Research studied how college students use Replika. The study observed, “Many used Replika in multiple, overlapping ways—as a friend, a therapist, and an intellectual mirror.”
Chatbots Galore
As ChatGPT demonstrates, it’s not just virtual companions becoming more human. Marketing and customer service chatbots are quickly evolving from functional to fun.
For instance, Startup Tymely is developing a chatbot to help brands offer email and chat support services in a more compassionate and accurate manner. Soul Machines has created a 3D digital individual currently being used on websites such as Calocurb. Users can choose to speak or type, and the digital individual will smile back if the camera is activated.
Chatbots are the go-to examples for brands because they represent a direct interaction between technology and users. They could be the brand’s first and only line of defense when people need to interact with a business. The more authentic and human a chatbot is, the more inclined we are to trust it emotionally, which is distinct from trusting it to deliver accurate information. Tone is crucial in human conversation and AI interactions.
Considerations for Brands
On the surface, who wouldn’t want a chatbot to be more human? However, brands need to tread carefully. One practical matter of sourcing content for voice styles carefully and thoughtfully, as OpenAI discovered when Scarlett Johansson threatened the company with legal action after a demonstration of GPT-4o sounded too much like her voice.
Businesses also need to align the tone of voice with the brand. AI should reflect the brand’s personality and values, maintaining a consistent tone of voice across all interactions while being adaptable to different contexts. For example, the Domino’s Pizza Dom chatbot helps users place orders, track deliveries, and answer common questions, with a personality that is efficient, friendly, and slightly playful. On the other hand, Spotify’s AI DJ feature curates personalized playlists and provides commentary in between songs with a cool, laid-back, and knowledgeable tone. Both approaches work for each brand and their purpose.
And let’s remember: AI still needs people. Brands need to maintain a level of human oversight to manage complex or sensitive interactions that the AI may not be equipped to handle. AI, like humans, makes mistakes. It’s a fast learner, though. With the right oversight, AI can build both emotional and intellectual trust that generates brand loyalty.
