AI risks - Revision history

Lê Nguyên Hoang: /* A list of risks */

2020-01-26T09:34:47Z

A list of risks

Lê Nguyên Hoang: /* Security mindset */

2020-01-24T18:19:27Z

Security mindset

Lê Nguyên Hoang: /* Security mindset */

2020-01-24T18:18:41Z

Security mindset

Lê Nguyên Hoang: Created page with "Numerours AI risks have been listed in books like [https://www.amazon.com/Superintelligence-Dangers-Strategies-Nick-Bostrom/dp/0199678111/ Bostrom14] [https://www.aaai.org/ojs..."

2020-01-20T21:23:26Z

Created page with "Numerours AI risks have been listed in books like [https://www.amazon.com/Superintelligence-Dangers-Strategies-Nick-Bostrom/dp/0199678111/ Bostrom14] [https://www.aaai.org/ojs..."

New page

Numerours AI risks have been listed in books like [https://www.amazon.com/Superintelligence-Dangers-Strategies-Nick-Bostrom/dp/0199678111/ Bostrom14] [https://www.aaai.org/ojs/index.php/aimagazine/article/download/2577/2521 RDT][https://dblp.org/rec/bibtex/journals/aim/RussellDT15 15] [https://www.amazon.com/Weapons-Math-Destruction-Increases-Inequality/dp/0553418815 ONeil16] [https://www.amazon.com/Life-3-0-Being-Artificial-Intelligence/dp/1101946598 Tegmark17] [https://www.amazon.com/AI-Superpowers-China-Silicon-Valley/dp/132854639X Lee18] [https://www.amazon.com/Army-None-Autonomous-Weapons-Future/dp/0393608980 Sharre18] [https://www.amazon.co.uk/Human-Compatible-AI-Problem-Control/dp/0241335205 Russell19] [https://laboutique.edpsciences.fr/produit/1107/9782759824304/Le%20fabuleux%20chantier HoangElmhamdi][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=Le+fabuleux+chantier%3A+Rendre+l%27intelligence+artificielle+robustement+b%C3%A9n%C3%A9fique&btnG= 19FR].

== A list of risks ==

They include cyberbullying [https://www.tandfonline.com/doi/pdf/10.1080/13811118.2010.494133?needAccess=true HindujaPatchin][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=Bullying%2C+Cyberbullying%2C+and+Suicide+hinduja+patchin&btnG= 10], fairness [https://arxiv.org/pdf/1908.09635 MMSLG][https://dblp.org/rec/bibtex/journals/corr/abs-1908-09635 19], privacy [https://link.springer.com/content/pdf/10.1007%2F978-3-030-30619-9.pdf TCK][https://dblp.org/rec/bibtex/conf/ml4cs/TanuwidjajaCK19 19], increased inequalities [https://www.sciencedirect.com/science/article/abs/pii/S0016328717300046 Makridakis][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=The+forthcoming+Artificial+Intelligence+(AI)+revolution%3A+Its+impact+on+society+and+firms&btnG= 17], job displacement [https://www.econstor.eu/bitstream/10419/202236/1/jrc-dewp201808.pdf MartensTolan][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=Will+this+time+be+different%3F+A+review+of+the+literature+on+the+Impact+of+Artificial+Intelligence+on+Employment%2C+Incomes+and+Growth&btnG= 18], radicalization [https://arxiv.org/pdf/1908.08313 ROWAM][https://dblp.org/rec/bibtex/journals/corr/abs-1908-08313 19], political manipulation [https://books.google.ch/books?hl=en&lr=&id=qTpxDwAAQBAJ&oi=fnd&pg=PP1&dq=political+manipulation&ots=fnNcRotnHb&sig=4pMUYcVS78JUYDkBKz352ChRcGo#v=onepage&q=political%20manipulation&f=false WoolleyHoward+][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=computational+propaganda+political+parties+politicians+and+political+manipulation+on+social+media&btnG= 18], misinformation [https://www.public.asu.edu/~huanliu/papers/Misinformation_LiangWu2019.pdf WMCL][https://dblp.org/rec/bibtex/journals/sigkdd/WuMCL19 19] [https://science.sciencemag.org/content/359/6380/1146/tab-pdf VRA][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=The+spread+of+true+and+false+news+online+vosoughi+roy+aral&btnG= 18], mute news (information that is drowned within the flood of information), information overload [https://link.springer.com/content/pdf/10.1007/s40685-018-0069-z.pdf Roetzel][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=Information+overload+in+the+information+age%3A+a+review+of+the+literature+from+business+administration%2C+business+psychology%2C+and+related+disciplines+with+a+bibliometric+approach+and+framework+developmen&btnG= 19], anger [http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.1019.3951&rep=rep1&type=pdf BergerMilkman][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=What+Makes+Online+Content+Viral%3F&btnG= 12], hate [https://www.aclweb.org/anthology/W17-1101.pdf SchmidtWiegand][https://dblp.org/rec/bibtex/conf/acl-socialnlp/SchmidtW17 17], geopolitical tensions, addiction [https://www.researchgate.net/profile/Ofir_Turel/publication/321472038_Time_distortion_when_users_at-risk_for_social_media_addiction_engage_in_non-social_media_tasks/links/5a24de660f7e9b71dd074688/Time-distortion-when-users-at-risk-for-social-media-addiction-engage-in-non-social-media-tasks.pdf TBB][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=Time+distortion+when+users+at-risk+for+social+media+addiction+engage+in+non-social+media+tasks&btnG= 18] [https://journals.sagepub.com/doi/pdf/10.1177/0894439316660340 HawiSamaha][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=The+Relations+Among+Social+Media+Addiction%2C+Self-Esteem%2C+and+Life+Satisfaction+in+University+Students&btnG= 16], inability to focus [https://dl.acm.org/ft_gateway.cfm?id=2858202&type=pdf MICJS][https://dblp.org/rec/bibtex/conf/chi/MarkICJS16 16], mental health [https://www.pnas.org/content/pnas/115/44/11203.full.pdf ESMUC+][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=Facebook+language+predicts+depression+in+medical+records&btnG= 18], loss of control [https://ora.ox.ac.uk/objects/uuid:17c0e095-4e13-47fc-bace-64ec46134a3f/download_file?file_format=pdf&safe_filename=Armstrong%2Band%2BOrseau%252C%2BSafely%2BInterruptible%2BAgents.pdf&type_of_work=Conference OrseauArmstrong][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=Safely+Interruptible+Agents+orseau+armstrong&btnG= 16], global (financial) crisis through unexpected disruption, resource depletion, autonomous weapons [http://faculty.engineering.asu.edu/acs/wp-content/uploads/2016/11/Ethics-of-Artificial-Intelligence-2015.pdf RHAV][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=Ethics+of+artificial+intelligence+russell+hauert+altman+veloso&btnG= 15], arms race [https://www.tandfonline.com/doi/pdf/10.1080/00963402.2016.1216672?casa_token=BzFIqiXh2-IAAAAA:PTd4zq-HQrrO8mmQ_yYzLqFpZhcfo06s4SjCGEhTGb0Hy875lRncWjleWuuHXlDIvP-I4PnCoFdaaA Geist][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=It%E2%80%99s+already+too+late+to+stop+the+AI+arms+race%E2%80%94We+must+manage+it+instead&btnG= 16], instrumental goals [https://www.aaai.org/ocs/index.php/WS/AAAIW16/paper/viewPDFInterstitial/12634/12347 BensontilsenSoares][https://dblp.org/rec/bibtex/conf/aaai/Benson-TilsenS16 16] and existential risk [https://www.researchgate.net/profile/James_Peters/post/Can_artificial_Intelligent_systems_replace_Human_brain/attachment/59d62a00c49f478072e9cbc4/AS:272471561834509@1441973690551/download/AIPosNegFactor.pdf Yudkowsky][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=Artificial+Intelligence+as+a+Positive+and+Negative+Factor+in+Global+Risk&btnG= 08] [https://onlinelibrary.wiley.com/doi/pdf/10.1111/1758-5899.12002?casa_token=q5x0kcZBAbsAAAAA:_oq2cid8PJNW9Lkb72tgkAnpaGbkQorLnFLcDabOx_s-AL98vin8NL4RkgcGd4TypbOHRBNiOgKm1JU Bostrom][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=Existential+Risk+Prevention+as+Global+Priority+bostrom&btnG= 13].

While there is already a lot of research, even more research on AI risks seems desirable, especially in areas with huge stakes and great uncertainty that could be reduced by more data collection or better thinking (see [[optimal exploration]]).

== Security mindset ==

It has been argued [https://www.lesswrong.com/posts/8gqrbnW758qjHFTrH/security-mindset-and-ordinary-paranoia Yudkowsky17] that there is a lack of security mindset, especially regarding performant large-scale algorithms. Security mindset can be regarded as focusing on worst-case (or near-worst-case) scenarios, especially if they are not completely unlikely and pose huge risks. This is in sharp contrast with Facebook's old motto "move fast and break things" [https://www.amazon.com/Move-Fast-Break-Things-Undermined/dp/0316275778 Taplin17]; and more generally with today's [https://en.wikipedia.org/wiki/Agile_software_development agile software development]. It is also in sharp contrast with current trial-and-error machine learning development, especially in light of [[Goodhart's law]] and the flaws of testing [work in progress].

More influential algorithms surely represent greater risks. Arguably, this is already evidenced by the case of [[YouTube]]. But crucially, we should not be [[overfitting]] on today's deployed algorithms to ponder risks from tomorrow's algorithms. Recall that the rise of machine learning algorithm is quite recent, and very spectacular. The security mindset urges us to prepare for a similar, if not much faster, rate of progress in the coming years. Not necessarily because this scenario is more likely. But because it is not completely unlikely, and poses much greater risks.

== Side effects ==

It is crucial to note that nearly all the above-mentioned AI risks are side effects of the large-scale deployment of algorithms (the main exception is autonomous weapons explicitly designed to cause harm). In particular, AI risks mostly don't come from the malicious intent of some algorithm, software developer or company. They mostly seem to arise from algorithms, software developers or companies' (sometimes intentional) tendency to neglect side effects of their behaviors.

To better understand this problem, it is worth comparing it to climate change. Greenhouse gas emission is not the intent of companies, car drivers and data centers. But it has become a problem because it was discarded as the companies', drivers' or data centers' goal and responsibility, sometimes intentionally, but also often by mere ignorance of the risks of such emissions.

The AI risk problem is similar. Unless we pay great attention to potential side effects and actively try to avoid them, the large-scale deployment of our algorithms will inevitably cause unintended potentially deadly side effects — as was the empowering of anti-vaccination propagandas (see [[YouTube]]).

Side effects of algorithms can also occur because of vulnerabilities of algorithms, typically to [[adversarial attacks]] [https://arxiv.org/pdf/1412.6572.pdf GSS][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=EXPLAINING+AND+HARNESSING+ADVERSARIAL+EXAMPLES&btnG= 15] [http://papers.nips.cc/paper/6617-machine-learning-with-adversaries-byzantine-tolerant-gradient-descent.pdf BMGS][https://dblp.org/rec/bibtex/conf/nips/BlanchardMGS17 17] [https://ijoc.org/index.php/ijoc/article/viewFile/6277/1804 NeffNagi][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=Talking+to+Bots%3A+Symbiotic+Agency+and+the+Case+of+Tay+neff+nagy+ijoc&btnG= 16].

Side effects were listed as the main concrete problem in AI safety by [https://arxiv.org/pdf/1606.06565 AOSCSM][https://dblp.org/rec/bibtex/journals/corr/AmodeiOSCSM16 16]. They are especially concerning for large-scale algorithms like social media recommender systems scholar. And they may be even more so, as such algorithms acquire more and more performant planning capabilities, built upon a much better modeling of their environments (see [[YouTube]], [[human-level AI]], [[AIXI]]).

In particular, algorithms with performant long-term planning capabilities feature risks of [[instrumental convergence]]. This corresponds to achieving instrumental goals like resource acquisition, which may endanger human activities or survival.

There seems to be a consensus that the only robust solution to AI safety and AI ethics is [[alignment]], typically based on [[volition]] learning [https://intelligence.org/files/CEV.pdf Yudkowsky][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=coherent+extrapolated+volition+yudkowsky&btnG= 04] and [[social choice]] aggregation [https://www.aaai.org/ocs/index.php/AAAI/AAAI18/paper/view/17052/15857 NGAD+][https://dblp.org/rec/bibtex/conf/aaai/NoothigattuGADR18 18]. A lot more research in this direction is needed.

@@ Line 13: / Line 13: @@
 More influential algorithms surely represent greater risks. Arguably, this is already evidenced by the case of [[YouTube]]. But crucially, we should not be [[overfitting]] on today's deployed algorithms to ponder risks from tomorrow's algorithms. Recall that the rise of machine learning algorithm is quite recent, and very spectacular. The security mindset urges us to prepare for a similar, if not much faster, rate of progress in the coming years. Not necessarily because this scenario is more likely. But because it is not completely unlikely, and poses much greater risks.
-The security mindset is argued to be hugely neglected by [https://nickbostrom.com/papers/vulnerable.pdf Bostrom18], especially given the current fast rate of discoveries and innovations. The paper also discusses global governance that may greatly increase or reduce risks caused by dangerous discoveries.
+The security mindset is argued to be hugely neglected by [https://nickbostrom.com/papers/vulnerable.pdf Bostrom18], especially given the current fast rate of discoveries and innovations. The paper also discusses global governance that may greatly increase or reduce risks caused by dangerous discoveries. This seems critical to [[AI governance]].
 == Side effects ==

← Older revision		Revision as of 18:18, 24 January 2020
Line 12:		Line 12:

	More influential algorithms surely represent greater risks. Arguably, this is already evidenced by the case of [[YouTube]]. But crucially, we should not be [[overfitting]] on today's deployed algorithms to ponder risks from tomorrow's algorithms. Recall that the rise of machine learning algorithm is quite recent, and very spectacular. The security mindset urges us to prepare for a similar, if not much faster, rate of progress in the coming years. Not necessarily because this scenario is more likely. But because it is not completely unlikely, and poses much greater risks.		More influential algorithms surely represent greater risks. Arguably, this is already evidenced by the case of [[YouTube]]. But crucially, we should not be [[overfitting]] on today's deployed algorithms to ponder risks from tomorrow's algorithms. Recall that the rise of machine learning algorithm is quite recent, and very spectacular. The security mindset urges us to prepare for a similar, if not much faster, rate of progress in the coming years. Not necessarily because this scenario is more likely. But because it is not completely unlikely, and poses much greater risks.
		+
		+	The security mindset is argued to be hugely neglected by [https://nickbostrom.com/papers/vulnerable.pdf Bostrom18], especially given the current fast rate of discoveries and innovations. The paper also discusses global governance that may greatly increase or reduce risks caused by dangerous discoveries.

	== Side effects ==		== Side effects ==

@@ Line 3: / Line 3: @@
 == A list of risks ==
-They include cyberbullying [https://www.tandfonline.com/doi/pdf/10.1080/13811118.2010.494133?needAccess=true HindujaPatchin][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=Bullying%2C+Cyberbullying%2C+and+Suicide+hinduja+patchin&btnG= 10], fairness [https://arxiv.org/pdf/1908.09635 MMSLG][https://dblp.org/rec/bibtex/journals/corr/abs-1908-09635 19], privacy [https://link.springer.com/content/pdf/10.1007%2F978-3-030-30619-9.pdf TCK][https://dblp.org/rec/bibtex/conf/ml4cs/TanuwidjajaCK19 19], increased inequalities [https://www.sciencedirect.com/science/article/abs/pii/S0016328717300046 Makridakis][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=The+forthcoming+Artificial+Intelligence+(AI)+revolution%3A+Its+impact+on+society+and+firms&btnG= 17], job displacement [https://www.econstor.eu/bitstream/10419/202236/1/jrc-dewp201808.pdf MartensTolan][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=Will+this+time+be+different%3F+A+review+of+the+literature+on+the+Impact+of+Artificial+Intelligence+on+Employment%2C+Incomes+and+Growth&btnG= 18], radicalization [https://arxiv.org/pdf/1908.08313 ROWAM][https://dblp.org/rec/bibtex/journals/corr/abs-1908-08313 19], political manipulation [https://books.google.ch/books?hl=en&lr=&id=qTpxDwAAQBAJ&oi=fnd&pg=PP1&dq=political+manipulation&ots=fnNcRotnHb&sig=4pMUYcVS78JUYDkBKz352ChRcGo#v=onepage&q=political%20manipulation&f=false WoolleyHoward+][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=computational+propaganda+political+parties+politicians+and+political+manipulation+on+social+media&btnG= 18], misinformation [https://www.public.asu.edu/~huanliu/papers/Misinformation_LiangWu2019.pdf WMCL][https://dblp.org/rec/bibtex/journals/sigkdd/WuMCL19 19] [https://science.sciencemag.org/content/359/6380/1146/tab-pdf VRA][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=The+spread+of+true+and+false+news+online+vosoughi+roy+aral&btnG= 18], mute news (information that is drowned within the flood of information), information overload [https://link.springer.com/content/pdf/10.1007/s40685-018-0069-z.pdf Roetzel][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=Information+overload+in+the+information+age%3A+a+review+of+the+literature+from+business+administration%2C+business+psychology%2C+and+related+disciplines+with+a+bibliometric+approach+and+framework+developmen&btnG= 19], anger [http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.1019.3951&rep=rep1&type=pdf BergerMilkman][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=What+Makes+Online+Content+Viral%3F&btnG= 12], hate [https://www.aclweb.org/anthology/W17-1101.pdf SchmidtWiegand][https://dblp.org/rec/bibtex/conf/acl-socialnlp/SchmidtW17 17], geopolitical tensions, addiction [https://www.researchgate.net/profile/Ofir_Turel/publication/321472038_Time_distortion_when_users_at-risk_for_social_media_addiction_engage_in_non-social_media_tasks/links/5a24de660f7e9b71dd074688/Time-distortion-when-users-at-risk-for-social-media-addiction-engage-in-non-social-media-tasks.pdf TBB][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=Time+distortion+when+users+at-risk+for+social+media+addiction+engage+in+non-social+media+tasks&btnG= 18] [https://journals.sagepub.com/doi/pdf/10.1177/0894439316660340 HawiSamaha][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=The+Relations+Among+Social+Media+Addiction%2C+Self-Esteem%2C+and+Life+Satisfaction+in+University+Students&btnG= 16], inability to focus [https://dl.acm.org/ft_gateway.cfm?id=2858202&type=pdf MICJS][https://dblp.org/rec/bibtex/conf/chi/MarkICJS16 16], mental health [https://www.pnas.org/content/pnas/115/44/11203.full.pdf ESMUC+][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=Facebook+language+predicts+depression+in+medical+records&btnG= 18], loss of control [https://ora.ox.ac.uk/objects/uuid:17c0e095-4e13-47fc-bace-64ec46134a3f/download_file?file_format=pdf&safe_filename=Armstrong%2Band%2BOrseau%252C%2BSafely%2BInterruptible%2BAgents.pdf&type_of_work=Conference OrseauArmstrong][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=Safely+Interruptible+Agents+orseau+armstrong&btnG= 16], global (financial) crisis through unexpected disruption, resource depletion, autonomous weapons [http://faculty.engineering.asu.edu/acs/wp-content/uploads/2016/11/Ethics-of-Artificial-Intelligence-2015.pdf RHAV][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=Ethics+of+artificial+intelligence+russell+hauert+altman+veloso&btnG= 15], arms race [https://www.tandfonline.com/doi/pdf/10.1080/00963402.2016.1216672?casa_token=BzFIqiXh2-IAAAAA:PTd4zq-HQrrO8mmQ_yYzLqFpZhcfo06s4SjCGEhTGb0Hy875lRncWjleWuuHXlDIvP-I4PnCoFdaaA Geist][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=It%E2%80%99s+already+too+late+to+stop+the+AI+arms+race%E2%80%94We+must+manage+it+instead&btnG= 16], instrumental goals [https://www.aaai.org/ocs/index.php/WS/AAAIW16/paper/viewPDFInterstitial/12634/12347 BensontilsenSoares][https://dblp.org/rec/bibtex/conf/aaai/Benson-TilsenS16 16] and existential risk [https://www.researchgate.net/profile/James_Peters/post/Can_artificial_Intelligent_systems_replace_Human_brain/attachment/59d62a00c49f478072e9cbc4/AS:272471561834509@1441973690551/download/AIPosNegFactor.pdf Yudkowsky][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=Artificial+Intelligence+as+a+Positive+and+Negative+Factor+in+Global+Risk&btnG= 08] [https://onlinelibrary.wiley.com/doi/pdf/10.1111/1758-5899.12002?casa_token=q5x0kcZBAbsAAAAA:_oq2cid8PJNW9Lkb72tgkAnpaGbkQorLnFLcDabOx_s-AL98vin8NL4RkgcGd4TypbOHRBNiOgKm1JU Bostrom][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=Existential+Risk+Prevention+as+Global+Priority+bostrom&btnG= 13].
+They include cyberbullying [https://www.tandfonline.com/doi/pdf/10.1080/13811118.2010.494133?needAccess=true HindujaPatchin][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=Bullying%2C+Cyberbullying%2C+and+Suicide+hinduja+patchin&btnG= 10], fairness [https://arxiv.org/pdf/1908.09635 MMSLG][https://dblp.org/rec/bibtex/journals/corr/abs-1908-09635 19], privacy [https://link.springer.com/content/pdf/10.1007%2F978-3-030-30619-9.pdf TCK][https://dblp.org/rec/bibtex/conf/ml4cs/TanuwidjajaCK19 19], increased inequalities [https://www.sciencedirect.com/science/article/abs/pii/S0016328717300046 Makridakis][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=The+forthcoming+Artificial+Intelligence+(AI)+revolution%3A+Its+impact+on+society+and+firms&btnG= 17], job displacement [https://www.econstor.eu/bitstream/10419/202236/1/jrc-dewp201808.pdf MartensTolan][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=Will+this+time+be+different%3F+A+review+of+the+literature+on+the+Impact+of+Artificial+Intelligence+on+Employment%2C+Incomes+and+Growth&btnG= 18], radicalization [https://arxiv.org/pdf/1908.08313 ROWAM][https://dblp.org/rec/bibtex/conf/fat/RibeiroO0AM20 20], political manipulation [https://books.google.ch/books?hl=en&lr=&id=qTpxDwAAQBAJ&oi=fnd&pg=PP1&dq=political+manipulation&ots=fnNcRotnHb&sig=4pMUYcVS78JUYDkBKz352ChRcGo#v=onepage&q=political%20manipulation&f=false WoolleyHoward+][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=computational+propaganda+political+parties+politicians+and+political+manipulation+on+social+media&btnG= 18], misinformation [https://www.public.asu.edu/~huanliu/papers/Misinformation_LiangWu2019.pdf WMCL][https://dblp.org/rec/bibtex/journals/sigkdd/WuMCL19 19] [https://science.sciencemag.org/content/359/6380/1146/tab-pdf VRA][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=The+spread+of+true+and+false+news+online+vosoughi+roy+aral&btnG= 18], mute news (information that is drowned within the flood of information), information overload [https://link.springer.com/content/pdf/10.1007/s40685-018-0069-z.pdf Roetzel][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=Information+overload+in+the+information+age%3A+a+review+of+the+literature+from+business+administration%2C+business+psychology%2C+and+related+disciplines+with+a+bibliometric+approach+and+framework+developmen&btnG= 19], anger [http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.1019.3951&rep=rep1&type=pdf BergerMilkman][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=What+Makes+Online+Content+Viral%3F&btnG= 12], hate [https://www.aclweb.org/anthology/W17-1101.pdf SchmidtWiegand][https://dblp.org/rec/bibtex/conf/acl-socialnlp/SchmidtW17 17], geopolitical tensions, addiction [https://www.researchgate.net/profile/Ofir_Turel/publication/321472038_Time_distortion_when_users_at-risk_for_social_media_addiction_engage_in_non-social_media_tasks/links/5a24de660f7e9b71dd074688/Time-distortion-when-users-at-risk-for-social-media-addiction-engage-in-non-social-media-tasks.pdf TBB][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=Time+distortion+when+users+at-risk+for+social+media+addiction+engage+in+non-social+media+tasks&btnG= 18] [https://journals.sagepub.com/doi/pdf/10.1177/0894439316660340 HawiSamaha][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=The+Relations+Among+Social+Media+Addiction%2C+Self-Esteem%2C+and+Life+Satisfaction+in+University+Students&btnG= 16], inability to focus [https://dl.acm.org/ft_gateway.cfm?id=2858202&type=pdf MICJS][https://dblp.org/rec/bibtex/conf/chi/MarkICJS16 16], mental health [https://www.pnas.org/content/pnas/115/44/11203.full.pdf ESMUC+][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=Facebook+language+predicts+depression+in+medical+records&btnG= 18], loss of control [https://ora.ox.ac.uk/objects/uuid:17c0e095-4e13-47fc-bace-64ec46134a3f/download_file?file_format=pdf&safe_filename=Armstrong%2Band%2BOrseau%252C%2BSafely%2BInterruptible%2BAgents.pdf&type_of_work=Conference OrseauArmstrong][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=Safely+Interruptible+Agents+orseau+armstrong&btnG= 16], global (financial) crisis through unexpected disruption, resource depletion, autonomous weapons [http://faculty.engineering.asu.edu/acs/wp-content/uploads/2016/11/Ethics-of-Artificial-Intelligence-2015.pdf RHAV][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=Ethics+of+artificial+intelligence+russell+hauert+altman+veloso&btnG= 15], arms race [https://www.tandfonline.com/doi/pdf/10.1080/00963402.2016.1216672?casa_token=BzFIqiXh2-IAAAAA:PTd4zq-HQrrO8mmQ_yYzLqFpZhcfo06s4SjCGEhTGb0Hy875lRncWjleWuuHXlDIvP-I4PnCoFdaaA Geist][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=It%E2%80%99s+already+too+late+to+stop+the+AI+arms+race%E2%80%94We+must+manage+it+instead&btnG= 16], instrumental goals [https://www.aaai.org/ocs/index.php/WS/AAAIW16/paper/viewPDFInterstitial/12634/12347 BensontilsenSoares][https://dblp.org/rec/bibtex/conf/aaai/Benson-TilsenS16 16] and existential risk [https://www.researchgate.net/profile/James_Peters/post/Can_artificial_Intelligent_systems_replace_Human_brain/attachment/59d62a00c49f478072e9cbc4/AS:272471561834509@1441973690551/download/AIPosNegFactor.pdf Yudkowsky][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=Artificial+Intelligence+as+a+Positive+and+Negative+Factor+in+Global+Risk&btnG= 08] [https://onlinelibrary.wiley.com/doi/pdf/10.1111/1758-5899.12002?casa_token=q5x0kcZBAbsAAAAA:_oq2cid8PJNW9Lkb72tgkAnpaGbkQorLnFLcDabOx_s-AL98vin8NL4RkgcGd4TypbOHRBNiOgKm1JU Bostrom][https://scholar.google.ch/scholar?hl=en&as_sdt=0%2C5&q=Existential+Risk+Prevention+as+Global+Priority+bostrom&btnG= 13].
 While there is already a lot of research, even more research on AI risks seems desirable, especially in areas with huge stakes and great uncertainty that could be reduced by more data collection or better thinking (see [[optimal exploration]]).