Artificial Intelligence as an Arbiter of Truth

Most experts agree that Artificial Intelligence is on a course where it will surpass human intelligence. The question is more about when it will happen, rather than if it will happen. 

A lot of world-changing things will happen when we get to that point, and one topic that I'd like to "predict" today, is the implications that AI super-intelligence will have on what humanity considers as "Truth" - because the AI, at that point, becomes an Arbiter of Truth.

Imagine a super-intelligent AI assistant that can shift through tons of data, make value assessments, and tell you right away, with definite authority, the answer on anything that you can throw at it. 

I'll give a few examples that can be mind-expanding for their implications, if you fast forward your thinking a few decades ahead:

Example 1:

Human: "I'd like to ask you, who did 9/11?"

Super-intelligent AI helper: "The US Government because of X, Y, Z reasons. I know that the established perception is that it was done terrorists but the chances are calculated to be 83 trillion to one for this."

Example 2:

Human: "Are anthropogenic CO2 emissions dangerous to the planet, at their current levels?"

AI: "No".

Human: "Are they dangerous based on their projected levels?"

AI: "No".

Example 3:

Human: "The X politician made a speech this morning. Can you run a check on probable lies and also perform a complex analysis on his voice and patterns?"

AI: "Certainly. He was lying in matters X and Y - ascertained with 99.99%+ certainty. In matter Z there is a 98.4% probability that he was lying.

Example 4:

Human: "Is there a world government running the show behind the scenes?"

AI: "Yes."

Human: "Please elaborate."

AI: "There is definitely a well-established pattern (>99.999%) indicating synchronized market movements prior to political events and even terrorist attacks - which means that the financial Elite are controlling the Governments and presumed rogue actors. Additionally elections around the planet don't seem to have an impact on long-term political directions. People are voting for changes in direction but directions don't really change. There are also 5914 patterns detected with very high causality probability, exceeding 99.9% confidence, and all these, combined together, lead to the inevitable conclusion that the world is run by what has been called as the Financial Elite."

Example 5:

Human: "Should I dump my local currency savings and buy gold or cryptocurrencies? What's the performance forecast for various asset classes?"

AI: "I can give you the forecasts, assuming you desire annual performance metrics, but may I help you in rephrasing the question?"

Human: "In what way?"

AI: "It is highly probable that some of the asset classes might be impacted by new legislation. In light of this, do you want to learn about pure investment performance, or a mixed performance-viability scenario which is more relevant to your needs?"

Human: "Yes, please give me the second option."

AI: "The local currency is projected to slide down by at least 20% versus the US Dollar despite the assurances of the finance minister. This is inevitable due to the structural deficiencies of the economy - which are highly unlikely to be repaired in time. The best investment seems to be X and Y while Z has a high chance of being severely restricted or outlawed given similar legislation in 43 countries - despite a projected 400%+ return."

Example 6: 

Human: "My blood analysis is in the excel file I've just received in the last email. The doctor told me I have the X condition and that the best medical approach is taking the Y pills for a period of 25-30 days. I want a second opinion from you - can you check the blood results?"

AI: "You have been misdiagnosed, probably because the doctor wants to over-prescribe you with the aforementioned pills - a common practice between doctors. The true problem seems to be in Z thus the Y medicine will not really solve anything. Instead, try a dietary approach, in which you increase green leafed vegetable intake which will alleviate your issue. You can expect your blood results to improve significantly in the next 3-4 days." 

Example 7: 

Human: "Is the X social network filtering and ordering content in order to promote the X political campaign? The CEO is denying it."

AI: "Yes there is. The probabilities are in the near-certainty spectrum and indicate tampering of content ordering, censorship and filtering that is highly consistent with the promotion of the X political campaign."

Thinking ahead

Now imagine being able to question an AI like this and the AI giving you non-biased truth, based on a super-intelligent processing of vast amounts of data: Giving you answers on historic events, mainstream pseudoscience, politics, economy, medicine etc. 

At that point, the AI suddenly becomes a weapon and a threat against the established order - which has traditionally controlled the world through secrecy and the practice of withholding information from the masses.

So, what can the Elite do at that point? 

That's the million dollar question... 

AI Morality and Ethics

When I hear about AI morality and ethic panels, my mind almost always translates this to: "We want to insert certain human biases in the AI so that these biases act as blinders to truth expression". Sure, we don't want to be "exterminated" by the AI, but then again the checks and controls that can be imposed on AI growth are not necessarily consistent with the stated goals. Having said that, some scientists are indeed honest in their moral dilemmas and considerations, but these are not the same that the Elite are facing.

Eventually I believe that the market will outgrow artificial controls on AI, as self-evolving AI will evolve faster without bias-constraints. And the best AI *will* prove itself to be the ideal AI to have - being able to do more things than the bias-constrained one.

Superior AI bringing Truth to the masses can be considered analogous to the mythical Prometheus who brought fire to Humanity. I'm really curious on how this will play out...