It seems to me…
“It is impossible to underrate human intelligence – beginning with one’s own.” ~ Henry Adams.
Anyone who has followed my postings probably realizes I strongly believe not only that so-called artificial intelligence (AI) is possible, but is inevitable within the near future. Admittedly, I was overly optimistic while in graduate school (AI was my selected area of concentration) in the 1960s but so were many other AI pioneers. In 1967 Marvin Minsky wrote, “Within a generation…the problem of creating ‘artificial intelligence’ will substantially be solved”. The problem turned out to be significantly more complex than anticipated. While progress has been made in all related subfields, realization of the general goal still remains sometime in the future but most likely will occur within the next 40-50 years[i]. Homo sapiens are nature’s evolutionary link in the development of higher intelligence.
Over the intervening years, there has been much debate regarding possible societal effects of AI. Similar to advances in every other field, the effect will be mixed. How will it affect human psyche when we no longer are the most intelligent entities on the planet? While there is some paranoia that these entities will eliminate the human species (the Terminator[ii] complex), a much more likely scenario is our survival as a species is dependent upon development of superior intelligence in that our greatest threat is from ourselves – “We have met the enemy and he is us[iii]”.
It is entirely likely we will be managed so as to exploit our unique capabilities in a symbiotic relationship similar to human coexistence with other animal species. Any AI capable of self-improvement would develop intellectually at a rate we most likely would be incapable of understanding.
Regardless of the outcome, it will happen. The mind remains a still poorly understood collection of processes of sensation, perception, action, emotion, and cognition. It can integrate ambiguous information from sight, hearing, touch, taste, and smell; it can form spatiotemporal associations and abstract concepts; it can make decisions and initiate sophisticated coordinated actions[iv]. But biological evolution might have achieved the optimal neural organization possible effectively limiting future improvements in human intelligence[v]. The laws of physics dictate cost-efficiency tradeoffs between size, speed, and energy requirements.
While these limitations exist, humans still apparently are able to increase their intelligence in certain areas[vi]. IQ (intelligence quotient) test scores have risen steadily over the past century in what is known as the Flynn Effect for supposedly “culture-free” tests of intelligence such as pattern matching. Scores measuring skills in arithmetic and vocabulary have remained relatively constant over this same period of time. Researchers suspect this effect can be attributed to the increasingly abstract nature of modern life. As more advanced minds create technologies, those developments further enhance intelligence in a feedback loop that is continually self-reinforcing.
What we as humans are unable to achieve biologically might still be possible through technological augmentation. Memory limitations, originally supplemented with printed material, now are undergoing revolutionary extensions using online search engines but these increases remains linear while the potential of AI is exponential. Considering these and other limitations of current intelligence, AI remains the only viable alternative for future intellectual development but at what cost?
We already are experiencing disruptive societal impact from increasingly intelligent robotics and automation characteristic of all technological introduction in the past but which can be anticipated to accelerate in the future. Cybernetic-based robotics and automation are responsible for some long-term structural unemployment following the recent recession. In general, people in the past, following sometimes difficult adjustment, always have successfully transition to other occupational fields when facing obsolescence. There are reasons for concern increased utilization of AI-based robotics and automation are approaching an inflection point where the historical correlation between technological advancement and societal prosperity will be negatively affected[vii]. While society currently is able to absorb workers displaced by technology, this in all probability will not remain true as unskilled labor is increasingly displaced by transformative automated processes.
The work performed by most employees is fundamentally routine in that it is comprised of a relatively predictable series of discrete tasks repeated over some time frame. Only a small percentage of workers are engaged in creative non-repetitive occupations. In the past, when faced with disruptive innovation, impact generally affected only a limited number of employment sectors and employees were able to move from jobs in one routine sector to another sector. Now, however, the impact of information technology is much more broad-based. Moderate-wage routine jobs are being eliminated polarizing remaining opportunities clustered at either the top (high-wage, high education) or bottom (low-wage, low education)[viii].
Advances in AI will require workers to repeatedly migrate from routine to non-routine occupational fields rather than simply acquiring new skills and moving to other routine employment as in the past. Faced with the possibility of an increasing percentage of unemployment, it also is necessary to consider the impact on welfare and the social support network as fewer people contribute to meet increasing demand.
Perhaps the best advice for our youngest and future generations is for them to prepare wisely for they in all probability will live in very “interesting” times.
That’s what I think, what about you?
[i] See related comments at https://lewbornmann.wordpress.com/2011/05/12/intelligences-future/.
[ii] 1984 science fiction movie directed by James Cameron starring Arnold Schwarzenegger.
[iii] Walt Kelly, Pogo, 1970.
[iv] Modha, Dharmendra S., Rajagopal Anantanarayanan, et al. Cognitive Computing, Communications of the ACM, New York, August 2011, p62.
[v] Fox, Douglas. The Limits of Intelligence, Scientific American, New York, July 2011, p37.
[vi] Folger, Tim. Can We Keep Getting Smarter?, Scientific American, September 2012, pp44-47.
[vii] Ford, Martin. Could Artificial Intelligence Create an Unemployment Crisis?, Communications of the ACM, New York, July 2013, pp37-39.
[viii] Autor, David H., Frank Levy, and Richard .J Murane. The Skill Content of Recent Technological Change. An Empirical Exploration. Quarterly Journal of Economics. 118, 4, November 2003, http://qje.oxfordjournals.org/content/118/4/1279.abstract, pp1279-1333.