Machines Will Be More Literate Than Many Humans in 10 Years

A global campaign has dedicated itself to bringing awareness that machines are getting more literate, while human literacy rates remain stagnant.

Are Siri, Cortana, and Alexa going to be more literate than humans?

Anyone excited about the recent advancements in artificial intelligence (AI) and machine learning should also be concerned about human literacy as well. That's according to Protect Literacy , a global campaign, backed by education company Pearson, aimed at creating awareness and fighting against illiteracy.

Project Literacy, which has been raising awareness for its cause at SXSW 2017, recently released a report, “ 2027: Human vs. Machine Literacy ,” that projects machines powered by AI and voice recognition will surpass the literacy levels of one in seven American adults in the next 10 years. “While these systems currently have a much shallower understanding of language than people do, they can already perform tasks similar to [a] simple text search task...exceeding the abilities of millions of people who are nonliterate,” Kate James, Project Literacy spokesperson and Chief Corporate Affairs and Global Marketing Officer at Pearson, wrote in the report. In light of this the organization is calling for “society to commit to upgrading its people at the same rate as upgrading its technology, so that by 2030 no child is born at risk of poor literacy.”

ESC Boston    is back! Get ready for an expanded, expert-led program over two full days, May 3-4, 2017 . Part of America's largest embedded systems industry conference series, ESC Boston gives you the chance to drill down into four tracks covering embedded hardware, embedded software, connected devices and the IoT, and advanced technologies.    Register Today

Compiling a variety of studies and reports, Project Literacy has highlighted some key findings:

According to the National National Centre for Education Statistics machine literacy has already exceeded the literacy abilities of the estimated 3% of non-literate adults in the US.

Comparing demographic data from the Global Developer Population and Demographic Study 2016 v2 and the 2015 Digest of Education Statistics finds there are more software engineers in the U.S. than school teachers, “We are focusing so much on teaching algorithms and AI to be better at language that we are forgetting that 50% of adults cannot read a book written at an eighth grade level,” Project Literacy said in a statement.

Research done by Business Insider reveals that 32 million Americans cannot currently read a road sign. Yet at the same time there are 10 million self-driving cars predicted to be on the roads by 2020. (One could argue this will further eliminate the need for literacy, but that is debatable.)

Citing research from Venture Scanner , Project Literacy found that in 2015 investment in AI technologies, including natural language processing, speech recognition, and image recognition, reached $47.2 billion. Meanwhile, data on US government spending shows that the 2017 U.S. Federal Education Budget for schools (pre-primary through secondary school) is $40.4 billion.

"Human literacy levels have stalled since 2000. At any time, this would be a cause for concern, when one in ten people worldwide...still cannot read a road sign, a voting form, or a medicine label,” James wrote in the report. “In popular discussion about advances in artificial intelligence, it is easy


Yeah . . . right. We see how well spell checker and grammar checker work. And someone thinks machines will be more literate? That's a good one.

What's the point of comparing machines to people who haven't been educated (or can't be educated for whatever reason)? If the point is to highlight the education problem, then AI and machines have nothing to do with that. Just write an article about addressing education issues. As for comparing machine intelligence to educated humans, we all know how well spell checker and grammar checker work.

Surpass the literacy levels of one in seven American adults ? That's a fairly low bar. It is still a long way from the level of surpassing the literacy levels of one in seven adults in general.

"One in seven American adults? Fairly low bar. " How fast do you think literacy is increasing in American adults? Let's assume it is increasing at an annual rate of 10%. How fast do you think literacy is increasing in machines? A conservative rate of 50% of the typical rate of Moore's Law would suggest machines literacy is doubling every 5 to 8 years. The implication is still unavoidable. The only debate is about the timeline.

My solution for increasing adult literacy would be quite effective, because it wold be able to deliver adequate motivation for everybody. Making two changes to the rules would be enough: First, Forbid the issuing of drivers permits and licenses to any illiterate, and, second, forbid any welfare payments to the illiterate. The main problems would be the large numbers seeking to learn, and the larger number of liberals complaining that somehow it is not fair.

As for machines becoming literate, that wold be OK as long as the computer machines do NOT also become self-aware. Consider the complications arising if a machine created to do dangerous jobs decided that it would not do them because they were dangerous. How would we deal with a robotic press unloading robot that refused to work because the task was unsafe?

Add new comment

By submitting this form, you accept the Mollom privacy policy.