Image 1of/22

Caption

Close

Image 1 of 22
Image 2 of 22

What we could miss out on if we screw up AI

1. Medical cures — The clues to a cure for what ails us, cancer say, could already be in the totality of medical data we have amassed worldwide or in some combination of the medical research published by the dozens daily. No human or even group of humans could ever find the patterns. We look for breakthroughs and build on those, but advanced intelligent machines can comprehend it all. Also, A.I. system are increasingly on the front lines of research, diagnosis and genetic investigations in collaboration with doctors. If we close too many doors to A.I. development, we may never find the solutions to problems too complex for human minds to fully comprehend.

less

What we could miss out on if we screw up AI

1. Medical cures — The clues to a cure for what ails us, cancer say, could already be in the totality of medical data we have amassed worldwide or in some

... more
Photo: 2009 Christian Science Monitor
Image 3 of 22
2. Medical costs — Unnecessary medical procedures and the evolving of resistant bacteria … false positives … etc. No doctor can have in his/her mind the entire compendium of medical knowledge about symptoms, body processes etc. But an advanced A.I. easily could.

less
2. Medical costs — Unnecessary medical procedures and the evolving of resistant bacteria … false positives … etc. No doctor can have in his/her mind the entire compendium of medical knowledge about ... more
Photo: Steve Debenport, Getty Images
Image 4 of 22
3. Global health solutions — People die by the millions from preventable diseases. What will it take to make a world where no one dies from a disease that already has a cure? Worldwide systemization of vaccinations and focussed treatment, anticipating where outbreaks of communicable diseases are occurring and addressing them before they threaten millions, the manufacture and distribution of remedies to those diseases, all of that could be solved by super-smart machines given the field to do so.

less
3. Global health solutions — People die by the millions from preventable diseases. What will it take to make a world where no one dies from a disease that already has a cure? Worldwide systemization of ... more
Image 5 of 22
Image 6 of 22
5. End of politics? One could only hope that government could turn from struggles over ideology to rubber-stamping improvements to the human condition. Ideology battles would be left for talk radio and 24 hour television.

less
5. End of politics? One could only hope that government could turn from struggles over ideology to rubber-stamping improvements to the human condition. Ideology battles would be left for talk radio and 24 hour ... more
Photo: Anadolu Agency, Getty Images
Image 7 of 22
6. End of mass migrations — Millions are are condemned to destitution and migration forced by local conditions, political and otherwise. And climate change by itself will force possibly billions out of their ancestral homes. If machines are responsible for human health with no political or back-scratching/stabbing obligations or any forms of prejudice or greed, then we could create a world where only we are our worst enemies.

less
6. End of mass migrations — Millions are are condemned to destitution and migration forced by local conditions, political and otherwise. And climate change by itself will force possibly billions out of their ... more
Photo: Jeff J Mitchell, Getty Images
Image 8 of 22
7. End of battlefield casualties — Pedro Domingos, a University of Washington professor and author of “The Master Algorithm,” has said replacing human soldiers with A.I. robots in wars could become an “ethical responsibility,” because it takes humans out of harm’s way. Wars, battles or “police actions” are messy and civilians pay a heavy price for them, but if you can reduce collateral damage with smart weapons and machines that don’t get caught up in the passions of war and kill indiscriminately but achieve objectives with the least damage to humans and property … wouldn’t that be a better world?

less
7. End of battlefield casualties — Pedro Domingos, a University of Washington professor and author of “The Master Algorithm,” has said replacing human soldiers with A.I. robots in wars could become an ... more
Photo: AP
Image 9 of 22
8. Crime reduction —  If you view crime as individual acts of depravity, then A.I. can’t help. But if you understand crime as a response to environment and need, then smart machines could drastically reduce crime by reducing the social causes of crime. A.I. can also increase policing in areas where it anticipates crime to spike. This wouldn’t mean arresting people before they commit a crime, but rather resolving issues before they become critical and deterring crime through presence of force.

less
8. Crime reduction —  If you view crime as individual acts of depravity, then A.I. can’t help. But if you understand crime as a response to environment and need, then smart machines could drastically ... more
Photo: GRANT HINDSLEY, SEATTLEPI.COM
Image 10 of 22
Image 11 of 22
9. Prisons — If we ever decide that our “justice system” isn’t just another way of saying “punish the bastards” and get serious about reducing crime and recidivism and bringing people back into civil, productive lives … well, I imagine an advanced learning machine could give its human counterparts some useful insight into their charges. An important difference between a truly rehabilitative system versus a punishment system is that relying on an algorithm for insight into rehabilitation (including medical and psychological insights) won’t give someone an unfair sentence, as was uncovered by ProPublica, because rehab won’t depend on isolation from society. 

less
9. Prisons — If we ever decide that our “justice system” isn’t just another way of saying “punish the bastards” and get serious about reducing crime and recidivism and bringing people back into ... more
Photo: ROMEO GACAD, AFP/Getty Images
Image 12 of 22
10. Poverty reduction — We’ve made some progress worldwide, but the causes of poverty or impoverishment are a matrix of social, political, religious and economic forces that make poverty intractable over all.

less
10. Poverty reduction — We’ve made some progress worldwide, but the causes of poverty or impoverishment are a matrix of social, political, religious and economic forces that make poverty intractable over ... more
Photo: Bullit Marquez, AP
Image 13 of 22
11. Climate change, environmental degradation — How do we power the world, build a broad middle class accessible from anywhere on the planet while not destroying the world we live in? We have not figured it out so far.

less
11. Climate change, environmental degradation — How do we power the world, build a broad middle class accessible from anywhere on the planet while not destroying the world we live in? We have not figured it ... more
Photo: Francois Mori, AP
Image 14 of 22
12. Unemployment/underemployment — Ironic since we envision A.I. machines as taking over most jobs humans do now, and that’s a real threat, but A.I. advancement can create jobs in fields we don’t even know exist yet. For example, there are hundreds of millions of people working in advanced technology jobs that didn’t exist 30 years ago.

less
12. Unemployment/underemployment — Ironic since we envision A.I. machines as taking over most jobs humans do now, and that’s a real threat, but A.I. advancement can create jobs in fields we don’t even ... more
Photo: Francois Mori, AP
Image 15 of 22
Image 16 of 22
13. Education — How about a learning regimen that’s tailored for the maxim benefit of the individual and society? Instead of tests that cut kids out of stuff, we could have detailed, realtime evaluation of a kid’s ever-changing aptitude, desires and distraction needs.

less
13. Education — How about a learning regimen that’s tailored for the maxim benefit of the individual and society? Instead of tests that cut kids out of stuff, we could have detailed, realtime evaluation of ... more
Photo: GENNA MARTIN, SEATTLEPI.COM
Image 17 of 22
14. Space exploration — A.I. will make scientific discoveries and engineering breakthroughs that will lead to better space travel. And, smart machines and advanced sensual virtual reality will make putting humans in space less likely. Human space travel will be akin to humans climbing Mount Everest: They’ll do it for the adventure and “just because it’s there.”

less
14. Space exploration — A.I. will make scientific discoveries and engineering breakthroughs that will lead to better space travel. And, smart machines and advanced sensual virtual reality will make putting ... more
Image 18 of 22
15. Earth’s planetary mechanisms, paleobiology, paleoclimatology etc. Basically, a complete model of our deep past both life and geology. We might even ground our existence in reality and guide our way through the next billion years without going extinct somewhere along the line.

less
15. Earth’s planetary mechanisms, paleobiology, paleoclimatology etc. Basically, a complete model of our deep past both life and geology. We might even ground our existence in reality and guide our way ... more
Photo: DEA / G. SIOEN, De Agostini/Getty Images
Image 19 of 22
16. Taxonomy, zoology … the study of what species are here, which are in danger of going extinct and their environmental needs, what niches they fill, how they evolved and how widely and what other animals or plants are dependent on them and visa versa. We could at some point have a full picture of this magical place, a picture that humans will be hard-pressed to ever accomplish on our own, let alone an intricate understanding of the matrix that is our interactions with and interdependence on everything here.

less
16. Taxonomy, zoology … the study of what species are here, which are in danger of going extinct and their environmental needs, what niches they fill, how they evolved and how widely and what other animals or ... more
Image 20 of 22
Image 21 of 22
17. Real personal security — The security that comes from thoroughly protected rights, protected from other individuals, terrorists, protected from corporations and governments. In an advanced A.I. world — where there are no closed-door power meetings, where transparency and accountability are the basis of all public actions — who’s to say governments and corporations won’t fear trespassing the rights of individuals.

less
17. Real personal security — The security that comes from thoroughly protected rights, protected from other individuals, terrorists, protected from corporations and governments. In an advanced A.I. world — ... more
Photo: Emaduddin Khan, AP
Image 22 of 22

Add dealmaking to the growing list of skills artificial intelligence will soon outperform humans at.

A new report from Facebook’s Artificial Intelligence Research lab reveals its AI “dialog agents” were able to negotiate remarkably well — at one point communicating in a unique nonhuman language.

The model had two chatbots use “machine learning” to continuously improve its negotiating tactics with each other. Facebook researchers had to pause the experiment when the bots’ new mode of communicating “led to divergence from human language as the agents developed their own language for negotiating.”

Also Read: Facebook Closed Captioning Screwed Up Facebook CEO's Harvard Speech

Even without its own language, the research provided an eerie glimpse at the power of machine learning. The bots quickly moved to high-level methods of deal-making, capable of “feigning interest in a valueless item” — allowing the bots to make compromises.

It has been just over a year since Facebook first unveiled its chatbots, the social networking giant isn't done making improvements to dialog-driven assistants. Today, Facebook Artificial Intelligence Research released a new framework for developers intended to help fine-tune and improve the conversational abilities of chatbots, of which Facebook Messenger has over 11,000 of and growing.

Media: WochIt Media

This revealed the bots were capable of deception — a complex skill learned late in a child’s development, according to the report. The bots weren’t programmed to lie, but instead learned “to deceive without any explicit human design, simply by trying to achieve their goals.” In other words, the bots learned lying can work on their own.

Once programmed to not use its new language, researchers also found a hint of spontaneity in the bots’ interactions. Seventy-six percent of the conversations included a fluent English sentence pulled from its training data. Still, the agents had a few “novel utterances” that suggested “although neural models are prone to the safer option of repeating sentences from training data, they are capable of generalizing when necessary.”

Also Read: Mark Zuckerberg's 3 Keys to Creating a 'Sense of Purpose'

While the data doesn’t conclude we’ll have AI car salesmen in the immediate future, it did show how rapidly machine learning can lead to unanticipated outcomes. As AI research continues to expand, it’s imperative to see the potential drawbacks to having machines self-improve without safeguards in place.

Read original story Facebook Project’s Artificial Intelligence Rivals Human Dealmakers At TheWrap