Behold, Intelligence! How Moving Beyond Mimicry Can Create Memorable AI

“Behold, a man!” This is what ancient Greek philosopher Diogenes exclaimed while holding up a plucked chicken to satirize Plato’s definition of a man as a “featherless biped”. The debate about what it means to be human is a topic that has fascinated humanity for millennia. Today, this debate is still taking place but, thanks to artificial intelligence (AI), it has taken a new dimension.

Ever since the creation of the “Logic Theorist”, experts have been trying to make machines and software imitate human skills and abilities. With AI now becoming more powerful than ever before, the debate about what it means to be human has been redoubled. Is “reasoning” all it takes for an application or machine to be intelligent? Is it creativity, empathy, or consciousness?

When generative AI models like ChatGPT were launched to the public, people all over the world were surprised at how “human” their output was. Trained using human-generated content, these models became very good at simulating human communication. However, to most experts, this is not enough to say these models possess intelligence.

When it comes to their code, generative AIs are nothing more than really good probabilistic models: they make predictions based on training data. They are impressive, sure, but they are not intelligent. The inflection point from these latest developments doesn’t come from these models being more intelligent than those created years ago but from their scale.

The information processed by models like GPT and Midjourney is massive, just as their performance at processing them. However, these models are unable to actually understand any concept or word they use, despite being able to describe them if asked. Neither GPT nor Google Search understand the information they are provided, they are only using data to provide you with an output. 

Yes, ChatGPT’s response might be much more complex and human-like than the results you get from Google. This human quality, however, is an illusion as it is only replicating patterns it has seen elsewhere. Some might make the case that this is the same humans learn and evolve but there is one key difference here: emotions, creativity, reasoning, and metacognition.

Now, AI is not about creating a new digital human in the sense of being a digital mirror of humanity in all of its aspects. As such, the types of approaches taken by generative and other types of AI are more than valid. However, understanding what these models are lacking and the difference between them being “intelligent” or “human” and mimicking a skill is quite important.

Most of the outrage and controversy around rapid AI development comes from a basic misunderstanding of the capabilities of generative AI. These people understand AI in the same way Plato did with its “featherless biped”, while critics are choosing to focus on the plucked chicken.

So how can these two mistakes be avoided? The easiest way would be to understand that intelligence is dynamic. Limiting the idea of intelligence to a few skills like Alfred Binet with the idea of IQ or adding more checklists is not the way to go. It is all about bringing multiple types of artificial intelligence together, allowing them to collaborate in a way akin to how humans do.

This is where the idea of “digital human” comes from, which relies on the understanding that human intelligence is complex and the result of many different skills. By bringing all of these together, no matter if they are “perfected” or not, something new and more complex emerges. 

UneeQ is an example of how generative AI can be imbued with empathy, personality, and other traits. The company has helped brands like L’oreal, Amazon Web Services, Dell, and Tmobile create digital humans who can not only help their customers but also empathize with them in a way reflective of their brands.

Danny Tomsett, CEO of UneeQ, talks about how his company is taking AI to the next stage by focusing on creating loving and breathing AI. He did so in a conversation with Forbes contributor, keynote speaker, and author Roger Dooley to talk about how focusing on “humanity” instead of “intelligence” could change AI forever. 

If you couldn’t attend this year’s edition of The Lively & Grit Daily House at SXSW 2024 and you would like to know what Tomsett had to say, you can watch the video below. For similar panels, make sure to check out Grit Daily’s YouTube channel.