Voice Assistants Learn the Art of Small Talk

Voice assistants have already made significant strides in areas such as smart home integration, educational settings and business applications. However, their current capabilities are limited by a lack of robust reasoning and planning abilities. 

Read more: AI’s First Year: How Voice Reshaped the Conversational Ecosystem

In fact, just 7.8% of consumers believe voice technology is as smart and reliable as a real person today, according to the PYMNTS Intelligence report “How Consumers Want to Live in the Voice Economy.”

This limitation is evident in the current state of the technology, which is predominantly device-driven and primarily used for simple tasks. For instance, while 52% of consumers have used voice commands to find and purchase airline tickets and accommodations on their mobile devices, only 44% completed the purchase entirely through voice prompts. 

However, that percentage could soon increase, with advancements in artificial intelligence (AI) paving the way for voice assistants that can complete more complex tasks.

Companies like Microsoft-backed OpenAI and Meta are spearheading this shift, pioneering the development of new AI models that promise to enhance voice assistants’ reasoning and planning capabilities.

As the Financial Times (FT) recently reported, OpenAI’s upcoming model, likely to be named GPT-5, is expected to address complex problems such as problem-solving, while Meta is rolling out Llama 3, an AI model embedded with reasoning, planning and memory capabilities.

“We’re going to start to see AI that can take on more complex tasks in a more sophisticated way,” OpenAI Chief Operating Officer Brad Lightcap told the FT. “I think we’re just starting to scratch the surface on the ability that these models have to reason.”

The integration of reasoning and planning into voice assistants is crucial for achieving what AI researchers term “artificial general intelligence” — superhuman cognition. This advancement allows voice assistants to comprehend and execute sequences of related tasks, predict outcomes and provide contextually relevant responses to user queries.

Enhanced reasoning and planning capabilities also pave the way for voice assistants to proactively assist users in accomplishing their objectives. For example, Meta is developing AI “agents” capable of planning and booking entire journeys spanning multiple destinations, highlighting the potential for voice technology to streamline complex tasks. 

The optimism surrounding the future of voice technology is evident among consumers, as the PYMNTS Intelligence study revealed that more than 60% of Americans believe that voice assistants will eventually match human intelligence and reliability in the near future. 

Millennials are the most optimistic, with 57% expecting this advancement within the next five years, surpassing the 46% average among all consumer demographics.

This optimism is further reflected in a willingness to invest financially in the promise of advanced voice assistants, with nearly 30% of American consumers open to paying a monthly fee for services that deliver the envisioned level of intelligence and reliability. Among millennials, this figure rises to 43%, reflecting a strong demand for next-generation voice technology among this age group. 

However, despite these promising prospects, challenges persist, including concerns about fraud, data privacy, security, algorithmic bias, accessibility and trust issues. Addressing these issues will be crucial to ensuring that the benefits of advanced voice assistants are responsibly managed and widely accessible to all users.