Asano underestimates AI
March 21. 2018 9:23PM
To the Editor: Sam Asano’s March 12 column suggests that Artificial Intelligence is essentially whatever computers do now, and cannot match the full set capabilities we associate with human intelligence. IEEE, the world’s largest society of technical professionals, has a standards effort to define ethics for artificial intelligence development, and many experts in the field have expressed serious concerns about the impact of unprincipled development in this area.
The majority of these experts anticipate artificial general intelligence (AGI), the technical term closest to what Sam describes, by 2047.
But this is all misleading as well. AI systems have gone past a tipping point in the last few years. The more narrow, “purpose built” AI systems are learning to win at games like Go, maximize gains in lighting stock market trading, manipulating consumers based on personalized psychographic analytics, surveillance of every citizen in China, and the range of focused social impacts is expanding exponentially, and at this time, not decades in the future.
The inability of AI systems to match (and then outperform) humans in every possible activity does not protect our jobs from robo-sourcing, our privacy from big data analytics, or social engineering of public opinion though personalized targeting.