Living human brain cells are training a chatbot to be 'more like us'

A company recently revealed its human-brain-cell-driven chatbot that it has taught to play video games.

However, even though the program runs on real human cells, it is still hallucinating answers.

'It could be more able to use biological intelligence in a meaningful way.'

The company, Cortical Labs, shared a video recently that showed its brain-cell-operated large language model responding in real time to a user prompt. Return reported on the company last year for using brain cells grown on a silicon "chip" for an organic computer. The tech is referred to as synthetic biological intelligence, and the company now appears to have a room full of such computers.

"This is a whole new paradigm to how ... the syntax of the LLM can be more like us," a company rep said in the video.

Boasting about his brain-powered chatbot, the voiceover added, "It could be more able to use biological intelligence in a meaningful way to select the next token and to create better responses."

This example immediately failed, though. When the user asked the program, "Tell me where you want to visit on vacation," the bot cited a place that does not exist.

"The Great Barrinchi Cove in the Maldives," it stated, before also suggesting "Tuscany, Italy, for its rolling hills, stunning views."

Despite the complex technology, these types of "hallucinations" are common with AI chatbots and are a somewhat typical occurrence for any frequent chatbot user.

RELATED: Dystopia: World's first 'biological computer' uses human brain cells that are 'raised in a simulation'

In 2022, Cortical Labs said it had successfully taught 800,000 living brain cells to learn how to play the game Pong. The cells were linked to a computer to gradually learn to sense the position of the game's ball and control the virtual paddle.

"If we allow these cells to know the outcome of their actions, will they actually be able to change in some sort of goal-directed way?" Chief Scientific Officer Brett Kagan asked at the time.

Since then, and after a bevy of requests, the company has advanced its model to play a more complex game: Doom.

The crude shooter game represents a 21-year jump in technological understanding for the program, given that Doom came out in 1993 and Pong in 1972.

RELATED: Trump fired Anthropic for being 'leftwing nut jobs,' but the company's AI is conquering the internet

The recent demo also showed an incomplete answer provided by the chatbot when asked to explain the meaning of life.

It answered, "The meaning of life is a philosophical question that has been pondered by thinkers and scholars across various cultures and religions for centuries."

It went on, "It explores the fundamental question of what purpose or ultimate meaning lies in human existence."

"Different interpretations may vary widely," it added before ending abruptly.

Cortical Labs has made its research publicly available, letting the viewer decide if the company is indeed creating science "for the greater good," as it has stated.

Like Blaze News? Bypass the censors, sign up for our newsletters, and get stories like this direct to your inbox. Sign up here!