Human Like Computers for Better Future

Artificial intelligence is often based on anthropomorphic comparisons. We are talking about artificial neural networks as “information processing paradigm that is inspired by the way biological nervous systems, such as the brain, process information”[i], and intelligent machines are often given human-like looks and / or characteristics.

We have already suggested that in order to render machine even more human-like, they could probably be tested with regards to their sense of humor. But what will make humans trust technology as they trust their colleagues, and what will be the work that intelligent algorithms will be in charge of? When analyzing the degree of trust that people have towards technology, it is surprising to see that in spite of the omnipresence of technology in human life, this degree is not very high. Thus Harvard Business Review article explains the concept of “algorithm avoidance” that consists in the fact that “people consistently prefer human judgment — their own or someone else’s — to algorithms, and as a result they often make worse decisions”[ii]. It also gives an example of the researchers from Northwestern’s Kellogg School and Harvard Business School interviewing members of the crowd sourcing site Mechanical Turk with regards to the tasks that could be outsourced to algorithms – those requiring “cognition” and “analytical reasoning” versus the ones that would require “feeling” and “emotion processing”. The result showed that when the task was described as requiring “feeling” and “emotion”, people had more difficulties attributing it to the machine than when it was described as requiring “analytical reasoning” because “it’s not OK for robots to feel things, because then they’re too close to being human”[iii]. Isn't it amazing that we are imitating human brain in order to create intelligent algorithms and making robots more humanoid looking but are afraid of the computers becoming “too human”?!

Still, making computers more autonomous and behaving like humans remains a big challenge for artificial intelligence researchers. There has been a vague of discoveries recently saying that the algorithms are taking on tasks that used to be reserved to humans previously and that the intelligent algorithms are now being given the emotional dimension which is “down to necessity”[iv]. The paradox is that we do not trust the machines 100%, and still “as everything around us, from phones to fridges, gets connected to the internet, we need a way to temper machine logic with something more human”[v].

Yet another challenge is undertaken by Google who works on the “algorithms with the capacity for logic, natural conversation and even flirtation”[vi]. An interesting thing about this is the concept of “thought vectors” – “a new type of algorithm designed to encode thoughts as sequences of numbers” and that will allow computers to “have common sense”[vii]. The algorithm is constantly learning by example and has been used for training Google Translate machine translation software. The “thought vectors” were used to extract the actual meaning of search queries, concepts standing behind the actual words “by ascribing each word a set of numbers (or vector) that define its position in a theoretical “meaning space” or cloud. A sentence can be looked at as a path between these words, which can in turn be distilled down to its own set of numbers, or thought vector. The “thought” serves as the bridge between the two languages because it can be transferred into the French version of the meaning space and decoded back into a new path between words. The key is working out which numbers to assign each word in a language – this is where deep learning comes in. Initially the positions of words within each cloud are ordered at random and the translation algorithm begins training on a dataset of translated sentences”[viii]. Examples in the article show a kind of ontological dependencies behind the notion of “thought vectors”: thus, “if you take the vector for Paris and subtract the vector for France and add Italy, you get Rome”[ix].

The direction of anthropomorphic developments in computation has been taken definitively and it seems to be the only way to allow trustful relationship and collaboration between humans and technology. In spite of the achievements in the field, there is still a way to go, as human workers are excelling in creativity, mobility and emotional perceptiveness[x]. Still, benefits from intelligent technologies will be perceivable if the social and economic infrastructure accompanies the technological development. One cannot agree with Erik Brynjolfsson and Andrew McAfee talking about five axes being education, infrastructure, entrepreneurship, immigration and basic research[xi].

As in the open letter signed by the researchers and businessmen, the focus should be on the economic and societal questions, so that the humanity could use intelligent technologies as a powerful tool for more efficient work, better jobs and a more interesting life.

Image source: pixabay.com

[i] Introduction to Neural Networks Using Matlab 6.0 by S. N. Sivanandam, S. N Deepa, Tata McGraw-Hill Education, 2006, p. 11

[ii] When Your Boss Wears Metal Pants by Walter Frick for Harvard Business Review, June 2015 issue, online https://hbr.org/2015/06/when-your-boss-wears-metal-pants?utm_campaign=Socialflow&utm_source=Socialflow&utm_medium=Tweet, accessed on May 25, 2015

[iii] When Your Boss Wears Metal Pants by Walter Frick for Harvard Business Review, June 2015 issue, online https://hbr.org/2015/06/w Rana el Kaliouby hen-your-boss-wears-metal-pants?utm_campaign=Socialflow&utm_source=Socialflow&utm_medium=Tweet, accessed on May 25, 2015

[iv] Rana el Kaliouby quoted in Say hello to machines that read your emotions to make you happy by Sally Adeen for New Scientist, May 14, 2015, online http://www.newscientist.com/article/mg22630212.900-say-hello-to-machines-that-read-your-emotions-to-make-you-happy.html#, accessed on May 23, 2015

[v] Rana el Kaliouby quoted in Say hello to machines that read your emotions to make you happy by Sally Adeen for New Scientist, May 14, 2015, online http://www.newscientist.com/article/mg22630212.900-say-hello-to-machines-that-read-your-emotions-to-make-you-happy.html#, accessed on May 23, 2015

[vi] Google a step closer to developing machines with human-like intelligence by Hannah Devlin for The Guardian, May 21, 2015, online http://www.theguardian.com/science/2015/may/21/google-a-step-closer-to-developing-machines-with-human-like-intelligence, accessed on May 22, 2015

[vii] Google a step closer to developing machines with human-like intelligence by Hannah Devlin for The Guardian, May 21, 2015, online http://www.theguardian.com/science/2015/may/21/google-a-step-closer-to-developing-machines-with-human-like-intelligence, accessed on May 22, 2015

[viii] Google a step closer to developing machines with human-like intelligence by Hannah Devlin for The Guardian, May 21, 2015, online http://www.theguardian.com/science/2015/may/21/google-a-step-closer-to-developing-machines-with-human-like-intelligence, accessed on May 22, 2015

[ix] Professor Geoff Hinton quoted in Google a step closer to developing machines with human-like intelligence by Hannah Devlin for The Guardian, May 21, 2015, online http://www.theguardian.com/science/2015/may/21/google-a-step-closer-to-developing-machines-with-human-like-intelligence, accessed on May 22, 2015

[x] The Great Decoupling: An Interview with Erik Brynjolfsson and Andrew McAfee, by Amy Bernstein and Anand Raman for MIT Technology Review, June 2015 issue, online: https://hbr.org/2015/06/the-great-decoupling, accessed on May 26, 2015

[xi] The Great Decoupling: An Interview with Erik Brynjolfsson and Andrew McAfee, by Amy Bernstein and Anand Raman for MIT Technology Review, June 2015 issue, online: https://hbr.org/2015/06/the-great-decoupling, accessed on May 26, 2015