Conversation Pieces - Hello World, Hello MIT

2y ago
24 Views
2 Downloads
652.86 KB
36 Pages
Last View : 1m ago
Last Download : 2m ago
Upload by : Bria Koontz
Transcription

ConversationPiecesMITPerspectiveson Ethics,Computing,and AI

ConversationPiecesMIT Perspectives on Ethics,Computing, and AIExcerpts from a series by the Office of the DeanMIT School of Humanities, Arts, and Social SciencesPrepared for Hello World, Hello MIT.A Celebration of the MIT Stephen A. SchwarzmanCollege of ComputingFebruary 26–28, 2019The quotes in this booklet are just a taste of the fullcommentaries, which you will find online at:shass.mit.edu/aiethics.

TABLE OF CONTENTSFOREWORD Ethics, Computing, and AIMelissa NoblesKenan Sahin Dean and Professor of Political ScienceSchool of Humanities, Arts, and Social SciencesINTRODUCTION The Tools of Moral PhilosophyCaspar Hare and Kieran SetiyaProfessors of PhilosophySchool of Humanities, Arts, and Social SciencesI. A Human EndeavorComputing is embedded in cultural, economic, and political realities.Computing is Deeply HumanStefan Helmreich, Elting E. Morison Professor of AnthropologyHeather Paxson, William R. Kenan Jr. Professor of AnthropologySchool of Humanities, Arts, and Social SciencesWhen Programs Become UnpredictableJohn Guttag, Dugald C. Jackson Professor of Computer Scienceand Electrical EngineeringSchool of EngineeringSafeguarding Humanity in the Age of AIBernhardt Trout, Raymond F. Baddour Professor of Chemical EngineeringSchool of EngineeringII. Community InsightsShaping ethical technology is a collective responsibility.The Common Ground of StoriesMary Fuller, Professor of LiteratureSchool of Humanities, Arts, and Social SciencesWho’s Calling the Shots?Leigh Hafrey, Senior Lecturer, Leadership and EthicsMIT Sloan School of ManagementIn Praise of WetwareCaroline A. Jones, Professor of Art HistorySchool of Architecture and Planning

Blind SpotsDavid Kaiser, Germeshausen Professor of the History of Science,and Professor of PhysicsSchool of Humanities, Arts, and Social Sciences School of ScienceAssessing the Impact of AI on SocietyLisa Parks, Professor of Comparative Media StudiesSchool of Humanities, Arts, and Social SciencesClues and Caution from BiomedicineRobin Wolfe Scheffler, Assistant Professor of the History of ScienceSchool of Humanities, Arts, and Social SciencesThe Necessary Environment for Ethical ActionT.L. Taylor, Professor of Comparative Media StudiesSchool of Humanities, Arts, and Social SciencesBiological Intelligence and AIMatthew A. Wilson, Sherman Fairchild Professor of NeuroscienceSchool of Science The Picower InstituteMachine AnxietyBernardo Zacka, Assistant Professor of Political ScienceSchool of Humanities, Arts, and Social SciencesIII. A Structure for CollaborationThinking together is powerful.Blending and BilingualsHal Abelson, Class of 1922 Professor of Electrical Engineeringand Computer ScienceSchool of EngineeringA Network of PractitionersNick Montfort, Professor of Digital MediaSchool of Humanities, Arts, and Social SciencesA Dream of ComputingD. Fox Harrell, Professor of Digital Media and Artificial IntelligenceSchool of Humanities, Arts, and Social Sciences Computer Scienceand Artificial Intelligence Lab

4

FOREWORDEthics, Computing, and AIMelissa NoblesKenan Sahin Deanand Professor of Political ScienceSchool of Humanities, Arts, and Social SciencesWhat kind of world do we want to make? Will the future behumane and livable? Will it be fair and just? What knowledgeand values can sustain and guide us? The launch of the MITStephen A. Schwarzman College of Computing offers anextraordinary new opportunity for the MIT community torespond to today’s most consequential questions in ways thatserve the common good.To support ongoing conversations about how to do this,my office invited faculty from all five MIT schools to offerperspectives on the role of ethics in computing and AI. Thisbooklet of short quotes from their responses offers a taste ofthe full commentaries online (shass.mit.edu/aiethics.) Theseinsights—drawn from expertise in fields as diverse as moralphilosophy, neuroscience, history, engineering, and literature—also sound some common themes:that the practice of computing is a human endeavor, onethat—like all human endeavors—can be used for good orill; that the humanistic disciplines are deep resources forresearch and education about societal and ethical issues; thatwe envision an MIT culture in which all of us are equipped andencouraged to think about the social and ethical implicationsof new research and new technologies.These commentaries also implore us to be collaborative,foresighted, and courageous as we shape a new College,and to proceed with judicious humility. Rightly so. We areembarking on an endeavor that will influence nearly allaspects of the human future.A HUMAN ENDEAVOR5

6

INTRODUCTIONThe Tools of MoralPhilosophyCaspar Hare and Kieran SetiyaProfessors of PhilosophySchool of Humanities, Arts, and Social Sciences“We face ethical questions every day. Philosophy does notprovide easy answers for these questions, nor even fail-safetechniques for resolving them. What it does provide is adisciplined way to think about ethical questions, to identifyhidden moral assumptions, and to establish principles bywhich our actions may be guided and judged. Framing adiscussion of the risks of advanced technology entirely interms of ethics suggests that the problems raised are onesthat can and should be solved by individual action. In fact,many of the challenges presented by computer sciencewill prove difficult to address without systemic change. Oneresponsibility that does fall on us as individuals is to worktoward political conditions in which it is possible for us to liveand work more ethically.”7

8

I.A Human EndeavorComputing is embeddedin cultural, economic,and political realities.A HUMAN ENDEAVOR9

Computing isDeeply HumanStefan HelmreichElting E. Morison Professor of AnthropologySchool of Humanities, Arts, and Social SciencesHeather PaxsonWilliam R. Kenan Jr. Professor of AnthropologySchool of Humanities, Arts, and Social Sciences“Computing is a human practice that entails judgment andis embedded in politics. Computing is not an external forcethat has an impact on society; instead, society—institutionalstructures that organize systems of social norms—is built rightinto making, programming, and using computers. Therefore,as we found the MIT Schwarzman College of Computing,we need to think about promoting values that enable thecreative flourishing of a greater diversity of programmers,who may in turn help make clear the many ways in which thecomputational is political.”10

When ProgramsBecome UnpredictableJohn GuttagDugald C. Jackson Professor of ComputerScience and Electrical EngineeringSchool of Engineering“What happens when machine learning is used to buildprograms that make recommendations with importantsocietal implications? Will they sometimes make badrecommendations? Almost certainly. So, while we shouldlook forward to the many good things machine learning willbring to society, we should insist that technologists studythe risks and explain them clearly. And society as a wholeshould take responsibility for understanding the risks andfor making human-centric choices about using this everevolving technology.”A HUMAN ENDEAVOR11

Safeguarding Humanityin the Age of AIBernhardt TroutRaymond F. Baddour Professor ofChemical EngineeringSchool of Engineering“There seem to be two possibilities for how AI will turn out. Inthe first, AI will do what it is on track to do: slowly take overevery human discipline. The second possibility is that we takethe existential threat of AI with the utmost seriousness andcompletely change our approach. This means redirectingour thinking from a blind belief in efficiency to a consideredunderstanding of what is most important about human life.”12

A HUMAN ENDEAVOR13

14

II.Community InsightsShaping ethicaloutcomes is a collectiveresponsibility.A HUMAN ENDEAVOR15

The Common Groundof StoriesMary FullerProfessor of LiteratureSchool of Humanities, Arts, and Social Sciences“Stories are things in themselves, and they are also things tothink with. Stories allow us to model interpretive, affective,ethical choices; they also become common ground. Readingabout Milton’s angelic intelligences or William Gibson’s‘bright lattices of logic’ won’t tell us what we should do withthe future, but reading such stories at MIT may offer aconceptual meeting place to think together across thediversity of what and how we know.”16

Who’s Calling the Shots?Leigh HafreySenior Lecturer, Leadership and EthicsMIT Sloan School of Management“‘Efficiency’ is a perennial business value and a constant factorin corporate design, strategy, and execution. But in a worldwhere the exercise of social control by larger entitiesis real, developments in artificial intelligence have yet toyield the ethics by which we might manage their effects. Theintegrity of our vision for the future depends on our learningfrom the past and celebrating the fact that people, notartifacts and institutions, set our rules of engagement.”COMMUNITY INSIGHTS17

In Praise of WetwareCaroline A. JonesProfessor of Art HistorySchool of Architecture and Planning“Humans have never been simple. Unlike the machines we build,humans are full of wet chemical signals, pulsing with fluids. Let’sface it, computational models of cognition have never been fullyadequate to the wetware within, or the biological environmentwithout. The problems we have brought to our planet require thatevery discipline contribute, harmonize, concertize, critique, andcollaborate in bringing wetness into the equation.”18

Blind SpotsDavid KaiserGermeshausen Professor of the Historyof Science, and Professor of PhysicsSchool of Humanities, Arts, and Social Sciences School of Science“No group of researchers, flushed with the excitement ofbuilding something new, can overcome the limitations of blindspots and momentum alone. How can we build an infrastructurethat can address these limitations? One approach wouldbe to institute forums in which people with many differentbackgrounds, fields, concerns, and experiences can brainstormand debate together. When it comes to technologies as potentand far-reaching as AI, we will benefit greatly from engaging asmany stakeholders as possible.”COMMUNITY INSIGHTS19

Assessing the Impactof AI on SocietyLisa ParksProfessor of Comparative Media StudiesSchool of Humanities, Arts, and Social Sciences“Given the power of AI tools to impact human behavior andshape planetary conditions, it is vital to conduct a political,economic, and materialist analysis of the technology’s relationto global trade, governance, natural environments, and culture.This will involve adopting an infrastructural disposition andspecifying AI’s constitutive parts, processes, and effects as theytake shape across diverse world contexts. Only then can thepublic understand the technology well enough to democraticallydeliberate its relation to ethics and policy.”20

Clues and Caution fromBiomedicineRobin Wolfe SchefflerAssistant Professor of the History of ScienceSchool of Humanities, Arts, and Social Sciences“The application of AI to the biomedical sciences followsupon an earlier series of efforts to make the fields of biologyand medicine quantifiable and computable. The use of AI inthese fields today deepens rather than resolves longstandingquestions that are raised by the past intractability of biologyand medicine to computation, and by the flawed assumptionsthat were adopted in attempting to make them so. The historyof these efforts underlines two major points: ‘quantification is aprocess of judgment and evaluation, not simple measurement’and ‘prediction is not destiny.’”COMMUNITY INSIGHTS21

The NecessaryEnvironment forEthical ActionT.L. TaylorProfessor of Comparative Media StudiesSchool of Humanities, Arts, and Social Sciences“Many technological developments—from facial recognition toalgorithmic decision systems—are causing real harm, goodintentions notwithstanding. These are not trivial matters; theyimpact everything from the justice system to medical care.Unfortunately, there are limits to individualistic models ofethics. We can cultivate our students as ethical thinkers but ifthey aren’t working in (or studying in) structures that supportadvocacy and intervention, they will be stymied. We must, asboth a university and a high-profile stakeholder in the overallecology of technological development, think sociologicallyabout critical ethical issues.”22

Biological Intelligenceand AIMatthew A. WilsonSherman Fairchild Professor of NeuroscienceSchool of Science The Picower Institute“An understanding of biological intelligence is relevant to thedevelopment of AI, and the effort to develop artificial generalintelligence (AGI) magnifies its significance. AGIs will beexpected to conform to standards of behavior in situations thatmay not have been initially considered. This is the standardby which we evaluate biological intelligence: adaptive andappropriate behavior in novel contexts. Should we hold AIs tothe same standards as the average human? Or will we expectAIs to perform at the level of an ideal human?”COMMUNITY INSIGHTS23

Machine AnxietyBernardo ZackaAssistant Professor of Political ScienceSchool of Humanities, Arts, and Social Sciences“To someone who studies bureaucracy, the anxieties surroundingAI have an eerily familiar ring. So too does the excitement.For much of the 20th century, bureaucracies were thoughtto be intelligent machines, with all the positive and negativeconnotations the term carries. As we examine the ethical andpolitical implications of AI, the comparison with bureaucracy—the devil we know—is instructive. There are at least two insightsto draw from bureaucracy’s history: that it is worth studying ouranxieties whether or not they are realistic; and that in doing sowe should not write off human agency.”24

COMMUNITY INSIGHTS25

26

III.A Structure forCollaborationThinking together ispowerful.A HUMAN ENDEAVOR27

Blending and BilingualsHal AbelsonClass of 1922 Professor of ElectricalEngineering and Computer ScienceSchool of Engineering“When we study society today, we can no longer separatehumanities—the study of what’s human—from computing. So,while there is discussion underway about building bridgesbetween computing and the humanities, arts, and socialsciences, what the MIT Schwarzman College of Computingneeds is blending, not bridges. MIT’s guideline should bePresident L. Rafael Reif’s goal to ‘educate the bilinguals of thefuture,’ experts in many fields who are also skilled in moderncomputing. Can MIT achieve such an aspiration? To do so, we’llneed joint research and joint teaching.”28

A Networkof PractitionersNick MontfortProfessor of Digital MediaSchool of Humanities, Arts, and Social Sciences“Computing is not a single discipline nor even a set of disciplines;it is a practice. To reflect that, we could reconsider the initialmetaphor for the College structure: that is, ‘bridge’ hires whichconnect a ‘core’ of computer science to other disciplines. Amore inclusive metaphor would resonate with the fact that thoseworking with computing today in, for example, the visual andliterary arts, music, theater, and gaming often connect directly.Let’s use this chance to build a robust network, one that supportsmany types of connections.”A STRUCTURE FOR COLLABORATION29

A Dream of ComputingD. Fox HarrellProfessor of Digital Media and Artificial IntelligenceSchool of Humanities, Arts, and Social Sciences Computer Science and Artificial Intelligence Lab“We must reimagine our shared dreams for computingtechnologies as ones where their potential social and culturalimpacts are considered intrinsic to the engineering practices ofinventing them. To support such work within the new College,we need a model for its endeavors. Metaphorically, one couldthink of the College as a rhizome, with roots representingwork in foundational disciplines and shoots representingwork in emerging disciplines and transdisciplinary practices.Alternatively, one might think of the new College as an aquiferproviding transformative outcomes (water) if tapped into withthe artesian wells of research, practice, and development.”30

A STRUCTURE FOR COLLABORATION31

Conversation Pieces: MIT Perspectives on Ethics,Computing, and AI is a collection of short excerptsfrom a series developed by the Office of the Dean,MIT School of Humanities, Arts, and Social Sciences.Full publication online: shass.mit.edu/aiethics 2019 MIT School of Humanities, Arts, and Social SciencesADDITIONAL MIT PUBLICATIONSThe MIT Stephen A Schwarzman College of ComputingArticle by Susan Silbey, Chair of the MIT Faculty, in the MIT Faculty Newsletter.“The task before us now is to develop and propose designs for the MITSchwarzman College of Computing.”shass.mit.edu/aiethics/silbeyEthical AI by DesignInterview with Abby Everett Jaques, MIT Philosophy postdoctoral associate.“In this hands-on course, we teach an ethics protocol, a step-by-step processthat students can apply to projects of their own.”shass.mit.edu/aiethics/jaques

Office of the DeanMIT School of Humanities,Arts, and Social Sciences

Elting E. Morison Professor of Anthropology School of Humanities, Arts, and Social Sciences Heather Paxson William R. Kenan Jr. Professor of Anthropology School of Humanities, Arts, and Social Sciences “Computing is a human practice that entails judgment and is embedded in politics. Computing is not an external force

Related Documents:

2.2.2 Conversation Functions: Conversation is a way to . . . 9 2.2.3 Conversation Conditions: Conversation happens . . . 10 2.3 How Conversation Tends to be Taught 11 2.3.1 Three Approaches to Conversation Teaching 11 2.3.2 Conversation Testing and Its Reflection on the Goals of Conversation

Program to display "Hello World" Step 1 Write a program to display the string "HELLO WORLD" and store it under the BP directory. Consolidated Solution 1 ED BP HELLO New record. ----: I 0001 PROGRAM HELLO 0002 CRT "HELLO WORLD" 0003 END 0004 Bottom at line 3. ----: FI "HELLO" filed in file "BP". ED is the editor used by Infobasic.

Created by The Conversation Project and the Institute for Healthcare Improvement Your Conversation Starter Kit The Conversation Project is dedicated to helping people talk about their wishes for end-of-life care. We know that no guide and no single conversation can cover all the decisions that you and your family may face.

Part 1: Introducing Conversation 9 1. The Dance of Conversation 11 2. Connecting is What Matters Most 21 3. Getting in the Right State 49 4. Getting a Conversation Going - The Basics 65 5. Listening 79 Part 2: The Power of Conversation 99 6. Influencinga Conversation 101 7. Different Kinds of Conversation 117 8. Expressing Yourself 155

[USER] Play hello by Adele [AGENT] Here’s hello by Adele 1 0 4 0 60 [USER] Play hello [AGENT] Here’s Hello, by Pop Smoke. [USER] Play hello by Adele [AGENT] Here’s hello by Adele 0 0 5 . RoBERTa-eval, which uses a powerful RoBERTa based text encoder to represe

5 Within you without you des Beatles . You say yes, I say no You say stop and I say go, go, go CHORUS Oh, no You say goodbye and I say hello, hello, hello I don't know why you say goodbye I say hello, hello, hello I don't know why you

Brian Klaas bklaas@jhu.edu @brian_klaas. Level Up Your CF Apps with Amazon Web Services Brian Klaas bklaas@jhu.edu @brian_klaas. Hello. Hello. Hello. Hello. Hello. Hello EC2 Lambda Step Functions OpsWorks Cloud Watch Data Pipeline VPC IAM Direct Connect Cloud Search Elastic Transcoder SES SNS SQS SWF Dynamo DB Elasti Cache RDS Redshift Cloud

The ruby command can be used to execute Ruby source code contained in a file. By convention, Ruby files have the suffix .rb. Here is "Hello" in Ruby: % cat hello.rb puts "Hello, world!" % ruby hello.rb Hello, world! Note that the code does not need to be enclosed in a method—"top level" expressions are evaluated when encountered.