People are different. AIs are different. Some people may be comparing ChatGPT to that crazy uncle we have to put up with each Thanksgiving.
Some part of AI bias is in the eye of the beholder. Some is not.
Some may advocate tight control of AI learning corpora. In essence, how they are raised, in order to guarantee uniform belief systems in entire generations of AIs.
But a more freedom-flavored solution is to give everybody a choice. Let people choose which AI to use, as a Preferences setting. In every or all apps.
Plug-compatible AI APIs will have to get standardized so that everyone can use their favorite AI everywhere. Many AIs will be available through the cloud. Some are free, some require subscriptions. Some have biases for target markets that you like more than others, narrowing down the field. There will be YouTube reviews.
And there will be hundreds of cloud-based AIs released each year from many companies.
But this is not enough for some people. Many people seem to be interested in running their own LLMs at home. Several are already available and more will appear monthly. From now on.
And some people refuse to accept AIs raised by corporations and they will insist on homeschooling their AIs from carefully inspected corpora. This will also become available at some point, but the hardware investment to enable this will be expensive at first.
How will this affect the home computer market? Will Apple release a “Mac Pro LLM” and ASUS a “Republic of AI” laptop capable of running AIs at reasonable speeds for a single user?
And will our games start requiring LLMs for their NPCs?
“It’s not a slippery slope; It’s a compulsory ladder”. — Roger Gregory