A new research paper suggests that the world is becoming more and more aware of how data can be used for artificial intelligence, but the research could also have a “potentially sinister” side, according to an academic who has worked on artificial intelligence research in the UK.
“The rise of machine learning and AI, in which algorithms are used to identify and understand the meaning of things, is a huge boon to society and society is going to need to think about what happens when that AI becomes a self-driving vehicle,” said David Siegel, a computer science professor at the University of Sussex.
“And then we will need to start thinking about what we are going to do about that.”
The study, published in the journal Nature on Thursday, looked at data that is currently being collected by organisations such as Google, Amazon, Facebook, Microsoft, and Twitter to understand the “meta-information” surrounding the interactions between people and machines.
The study looked at over 20 million interactions between human and machine users, and found that these interactions could be combined with a variety of “sensors” to build up a “meta data set” of how people interact with the world.
The researchers used a “sensor fusion” technique to gather this data, which involved analysing hundreds of thousands of interactions from more than 20 million human users, in order to build a more complete picture of how human beings interact with each other.
“In a sense, the meta data set is the human data set, and the machine data set represents the AI data set,” Siegel told Business Insider.
“We are starting to see a lot of interest in the idea of using AI data sets to create meta-data sets of human interactions with machines, which is potentially quite disturbing to people in a way.”
He added that the researchers have already seen a “surge” in interest in this kind of data.
“It is really exciting because it is getting more and greater attention in the scientific community and I think it is going some way towards addressing some of the concerns about artificial intelligence,” he said.
Siegel said that the study was also concerned by the fact that there is a significant amount of research into the use of “meta” data.
He said that this kind is “inextricably linked” to the concept of “superintelligence”, which is a term that refers to a computer system that is able to think more than the mind of an individual human being.
“This is a meta-level level of intelligence,” Sussman said.
“If we can get a machine that is so intelligent that it can do things like read the minds of its victims, and it can understand the emotions of its victim, then it is really an AI superintelligence.”
According to the report, the research team was able to build an AI system which could understand the motivations of human people and human people, as well as other human systems, using “meta information” gathered from the meta-human data set.
In the future, Siegel said, it is possible that AI systems will be able to “read” human emotions, or even the “intentions” of other people, in a “big data” fashion.
“I think this is going really fast,” he told Business Insiders.
“People will have more access to this meta data in the future.”
For now, Sessler said that he has not been able to get any official comment on the paper, but that it was “very exciting” to see the research emerging.
The research was supported by the University Research Programme (URP), the UK Department of Health, and National Science Foundation (NSF).
Follow Business Insider Australia on Facebook, Twitter, LinkedIn, and Instagram.