Ethical Artificial Intelligence: Public Perception and Bias (Current…
Ethical Artificial Intelligence: Public Perception and Bias
current status of AI and bias
AI is currently being used in inappropriate ways to fill human, social roles. These uses are hurting people, specifically minorities.
Several current and popular implementations of AI have ultimately failed because they were biased in some way.
AI technologies are already beginning to fill social roles in which bias is not acceptable such as character-assessment or facial-recognition.
Currently AI is being used to fill social roles that were once filled by humans that were vetted and scrutinized. For example, an algorithm that asses the likelihood of a criminal to commit another crime or facial recognition that allows someone to enter their apartment building.
Part of why AI is biased is that these algorithms are biased and discriminatory against minorities because they are being built and tested by a very small and homogenous group of people.
Bias is a key piece of the ethical puzzle around AI.
Current perception of AI
Currently, the general public has an overwhelmingly negative attitude toward AI.
Understanding these fears better in order to understand how we can implement AI in a safe and comfortable way moving forward.
Despite recent controversies around ineffective, biased, algorithms the government and tech industry aren’t doing enough to improve their algorithms and avoid these issues moving forward.
AI isn’t something of the future, it’s being used right now and is likely affecting your life in some way.
Even if current, biased algorithms aren’t directly putting you at risk, they are hurting a fellow member of your society.
This negative reception is often attributed to negative portrayals of AI in media, or a general mistrust in technology.
AI's current failures due to bias are damaging AI's perception. Before AI can actually do good for our society it has to be accepted.
The need for ethical AI
If AI has the potential to reach superhuman levels we want to try our best to ensure that entities are as ethical, unbiased, and just as possible.
The optimistic angle on AI is that is has the ability to help humanity tackle complex issues like cancer and water scarcity.
AI has the potential for extreme good, but I don’t think that will happen naturally, we have to work really hard to make sure that happens.
AI is only going to become more powerful and more commonplace. Everyone should care about AI being as ethical and fair as possible because it will soon have significant influence over our lives.
Many theorists posit that superhuman AI is inevitable. If something is going to be more powerful than humanity we should want it to be ethical.
Biased algorithms are just the tip of the iceberg, what happens when an algorithm is biased and more powerful than humans?
working toward ethical AI
We have the power to encourage our government and tech industry to place more consideration onto the ethics of AI.
I hope that understanding the central role it will soon play in our lives will motivate people to want to understand how AI can be more ethical and what we can do about it as individuals.
Bostrom asserts that superhuman intelligence will require superhuman ethics if we actually want to feel safe/comfortable in the presence of future AI.
The government has the ability to establish laws around how AI is made and tested. These laws could prevent bias and also misuse.
Bostrom also argues for transparency and predictability for AI that fills social roles.