In nature, the foundation of intelligence for animals, humans and even plants is based on objective data such as neurotransmitter chemical signals, heart rate, sun intensity, etc. This type of data is gathered through observation and measurements regardless of the feelings and opinions of the person selecting, presenting, or receiving the data. Nature has well-established protocols to parse and act on this information.
On the contrary, with Artificial Intelligence, algorithms are not built as protocols and data is not objective, rather it is subjective. A large part of what we do is capturing, interpreting, cleaning, labeling, and maintaining data that we use to train our models. This subjective approach is prone to producing inconsistent outcomes.
In our quest to build Artificial Intelligence systems we should aspire to build algorithms and models that are consistent, truthful and secure.
Objective data and protocols are the foundation for how intelligence exists in life and should be an aspiration in our approach. The ghost in the machine is not so much of a ghost, but is instead created from our own data selection bias.
Dr. Julieta Ross 