‘OI’ Can Overtake AI

Recently, it has seemed as though artificial intelligence was taking over our lives. But, a group of scientists believes that “organoid intelligence,” or OI, which is driven by living human brain cells, may one day outperform all artificial systems and execute tasks far more effectively.

Organoids are clumps of biological tissue that are three dimensional, and scientists have been developing and experimenting with them for years. In order to create a “biocomputer” powered by human brain cells, Johns Hopkins University researchers under the direction of Thomas Hartung, professor of environmental health sciences, are experimenting with brain organoids.

The technology needed to implement true biocomputing have reached a point in development, according to Hartung.  The hope is that some of the extraordinary functions of the human brain can be implemented as OI, such as its capacity for making snap judgements based on conflicting and inadequate information.

In the Tuesday issue of Frontiers in Science, Hartung and colleagues present a comprehensive vision for the future of OI.

The group includes researchers from Cortical Laboratories, which made news last year when it developed a dish full of living brain cells that swiftly learned how to play the first Pong computer game.

For scientists, using organoids made from cells is useful since it eliminates the need for testing on humans or animals. Using human skin cells that have been reprogrammed into an embryonic stem-cell similar state, Hartung has been producing functional brain organoids since 2012. After that, they can be utilized to create brain cells and, eventually, organoids with living neurons and other characteristics that can maintain essential abilities like memory and ongoing learning.

The speaker responds, Because you can start gaming the system and doing things that are morally terrible to do with human minds.

A biological computer

He and his colleagues anticipate combining brain organoids into new biological computing technology that would be significantly more energy-efficient than supercomputers today.

Modern computers are still unable to match the brain, according to Hartung. “A 6,800 square foot installation costing $600 million, Frontier is the newest supercomputer in Kentucky. It only surpassed the processing power of a single human brain for the first time in June of last year, albeit at the cost of a million times more energy.”

Hartung acknowledges that computers handle data and numbers more quickly, but he still believes that brains are superior to machines at solving difficult logical puzzles.

Even though scientists strove to make computers more resemble the brain from the beginning of the computer age, they are not the same. OI promises to add a few more traits.

A library’s worth of new ethical debates could be sparked by ideas like biological computers and organoid intelligence. Although though the technology is believed to be in its infancy at the moment, discussions about organoids becoming sentient, conscious, or self-aware and the following ramifications have been ongoing for years.

According to Hartung, there probably isn’t any technology without unintended repercussions. Even if it is difficult to completely eliminate such hazards, humans still have control as long as they are in charge of input, output, and the feedback that goes to the brain regarding the effects of its output. The issue arises, though, the moment we grant AI or OI autonomy. Machines  that are built on cellular or siliceous machinery cannot make decisions regarding human life.

To evaluate the ethical ramifications of using OI, members of the research team with backgrounds in bioethics have been working.

AI and human brains grown the traditional manner won’t be in danger any time soon from organoid intelligence or biocomputers. To overcome some of the limitations of current silicon systems, though, Hartung thinks it’s time to start boosting the manufacturing of brain organoids and training them using AI.

The goal of developing anything comparable to any kind of computer will take decades to realize, according to Hartung. It will be much more challenging if they don’t start developing funding programs for this, though.

Source link

Most Popular