AI presents big challenges

King Trump has found another issue to dismiss as a hoax. First, it was the investigation into Russia’s role in the 2016 election. Then it was climate change, which he has dismissed as a hoax perpetuated by China to turn the American public against the consumption of fossil fuels. Next it was the issue of “affordability” dismissed by Trump as a hoax invented by Democrats to distract the American voters from his great accomplishments during his second term. And now it is the call for a slowdown and an expressed need for safeguards before proceeding with additional experimentation with artificial intelligence models which have earned the hoax label from Trump.

Trump has stated that the recent safety concerns expressed by multiple AI leaders are the work of conspiracy theorists. He added it is a “sick conspiracy” and the only control needed is a “STRONG and SMART (High IQ) PRESIDENT” (Trump’s capitalization) which we apparently enjoy “in spades”. He argued we must not pause AI experimentation because China would gain an advantage and exclaimed “Whoever wins AI, wins!”

The most recent “hoax” issue regarding a needed slowdown and the introduction of safeguards prior to proceeding with additional AI model experimentation is particularly complex and Trump’s opposition to the slowdown is not without some merit. The AI industry is a significant part of our economy and is at the center of the substantial stock market growth investors are enjoying. The industry promises to reduce the expense and increase the speed of completing complex tasks, predictions and computations in a wide range of human endeavors resulting in a better quality of life for humanity. In the US, the enterprise has emerged out of the private sector and has progressed rapidly with few government safeguards or regulations during this period of growth. Up to this time, the application of AI has served primarily civilian enterprises. This pattern of growth has started to change as military demands for AI applications increase and the fear of a renegade AI agent directing atomic or conventional weapons deployment is expressed. Fears beyond human loss of control include massive misinformation dissemination, disastrous cyberattacks, universal surveillance resulting in loss of privacy, loss of employment and a dulling of the human mind and skill sets because of overreliance on AI.

Another set of AI issues revolves around the massive data center infrastructures necessary to support the unlimited data storage and data application associated with the AI industry. Citizen opposition to the centers is strong because of tremendous water usage and energy consumption for plant operations, as well as other environmental and community disruption concerns. A final set of AI related issues is found in the incredible wealth being generated by the industry for a limited few entrepreneurs. These few oligarchs are now asserting undue influence throughout our democracy and threaten the very integrity of our system of governance. Citizens rightfully ask if the net value to society through development of artificial general intelligence or the more ambitious artificial super intelligence outweighs all the problems created by the endeavor.

The public demand for AI safeguards will ultimately result in more governmental intervention as the public is skeptical of allowing the lead actors in the enterprise to regulate themselves. The reputation of the few billionaire oligarchs leading the industry as having an overzealous desire to maximize profits provides good reason to question their commitment to serving the public good. It is difficult to project the governmental mechanism that might be put in place to regulate AI and the specific regulations that might emerge. Regulation supporters have expressed support for independent parties equipped with a “kill switch” to control worst-case scenarios and independent evaluators embedded in AI operating models. Congress is without the expertise to regulate directly but a body of civilian AI experts comparable to the Atomic Energy Commission might be created to promulgate a body of regulations and to litigate controversies arising from the regulations.

The fear that our decision to regulate the AI industry will provide other nations an advantage and foes will turn their advancements against us needs to be addressed. We had similar concerns as we introduced safeguards for use of atomic weapons and regulations impacting the development of nuclear energy in the private sector. The resolution of our fear that we would be taken advantage of by other nations came with the signing of the Nuclear Non-Proliferation Treaty in 1968. The treaty was extended indefinitely in 1995 and currently includes the signatures of 191 nations. Israel, India and Pakistan did not sign the treaty and they have developed atomic weapons. North Korea signed the treaty in 1985 but withdrew in 2003 and conducted underground nuclear weapon tests in 2006. Fears that Iran is on a path to develop an atomic weapon has been stated by Trump to be one of the reasons for our current war with Iran. With the exceptions noted above, the Nuclear Non-Proliferation Treaty has succeeded in deterring nations from developing nuclear weapons with the understanding that other nations will abide by the same rules. An international treaty like the Nuclear Non-Proliferation Treaty restricting certain military and civilian application of AI might succeed in deterring the reckless and life-threatening development and application of AI models. It is unlikely Trump and the Senate will work together to advance such a proposal with other nations, but we can trust future leaders will behave in a more rational manner.