Elon Musk, co-founder of Tesla, SpaceX and other companies, has said that Artificial Intelligence (AI) is the most serious threat to humanity. At an MIT event in 2014, Mr. Musk noted, “I think we should be very careful about artificial intelligence. If I had to guess at what our biggest existential threat is, it’s probably that. So we need to be very careful. I’m increasingly inclined to think that there should be some regulatory oversight, maybe at the national and international level, just to make sure that we don’t do something very foolish.”
The challenges that Mr. Musk states indicate that regulation is needed, both at national and international levels, and that care is needed in the implementation of AI. A quick rush to implement algorithms, to improve Machine Learning (ML) and to use AI for a competitive advantage has great potential to open AI up to unsuitable uses or fields. In addition, there are unforeseen consequences – things that the creator of the algorithm or Machine Language did not or could not foresee, that could be dangerous or even disastrous.
As Cathy O’Neil noted in her book Weapons of Math Destruction, the dangers of rapid and unthinking adoption of ML and AI in a variety of fields conclude that predictive models are never neutral, and reflect the goals and ideologies of the creator of the model. Models tend to favour the well-to-do and educated. Since the AI algorithms are based on predictive models, and the models are written by humans, there is the highly likely potential to insert bias, even unintentional. Models must be designed and tested to ensure that bias is minimized, and potentially reduced by the Machine Learning.
If AI standards can be written and agreed upon, then the next question is how to regulate the AI. Blockchain technology provides a tamper-proof record ledger, and, as such, can provide the immutable tamper-proof record of actions for AI regulation. In addition, Blockchain can enable access control based on reputation, and only allow reputable autonomous AI entities access.
Once the standards are defined, AI would write to a Blockchain and access controls would monitor the reputation of the entity. The controls in place would ensure that only reputable entities write changes, and also ensure that the standards are met.
In Thomas Hobbes’ text, Leviathan, he wrote that citizens must give up some of their power to a sovereign authority or Leviathan, in order to maintain a societal contract with others. Creating the Leviathan, however, can leave some of the power to a fallible human. If used in programming AI, the leviathan Blockchain can enforce identity, reporting, and manage the standards. Perhaps, Blockchain can save humanity.