Doctoral position in artificial intelligence and political language with the GRIPES project

Dogwhistles are political messages that are optimized to be understood with their original intent by some groups in the population but not others – a dogwhistle is “defeated” if everyone understands it. Dogwhistles are used to propagate publicly “unacceptable” opinions, such as racist or anti-scientific viewpoints, and deliver messaging advantages to the political factions that are using them.

The current state of AI technology and isolated scandals (e.g., the Cambridge Analytica affair) suggests that well-resourced private and state actors may be doing this kind of optimization behind closed doors using large computational resources. GRIPES seeks to understand the limits of these technologies and develop tools in the public interest to make the role of AI in political messaging more transparent.

The University of Gothenburg conducts leading research in computational linguistics, theoretical linguistics, journalism, media and opinion research through multiple units across faculties. GRIPES is hosted by the Department of Philosophy, Linguistics, and Theory of Science (FLoV) at the University is closely associated with the Centre for Linguistic Theory and Studies in Probability (CLASP). It also has participation from the Society-Opinion-Media (SOM) Institute and the Department of Journalism, Media, and Communication (JMG). GRIPES is also part of the WASP-HS program, the humanities and social sciences section of the Wallenberg Autonomous Systems Program.

This PhD position is part of the Wallenberg AI, Autonomous Systems and Software Program on Humanities and Society (WASP-HS). WASP-HS aims to realize excellent research and develop competence on the consequences and challenges of artificial intelligence and autonomous systems for the individual person and society. This 10-year program is initiated and generously funded by the Marianne and Marcus Wallenberg Foundation (MMW) with 660 million SEK. In addition to this, the program receives support from collaborating industry and from participating universities.

WASP-HS includes an extensive national graduate school with up to 70 doctoral students, the creation of least ten new research groups across Sweden, support for twelve visiting professors to strengthen Swedish research and networking activities, and a number of research projects. For more information about the research and other activities conducted within WASP-HS, please visit http://wasp-hs.org/.

The WASP-HS graduate school provides foundations, perspectives, and state-of-the-art knowledge in the different disciplines taught by leading researchers in the field. Through an ambitious program with research visits, partner universities, and visiting lecturers, the graduate school actively supports forming a strong multi-disciplinary and international professional network between PhD students, researchers and practitioners in the field. It thus provides added value on top of the existing PhD programs at the partner universities, providing unique opportunities for students who are dedicated to achieving international research excellence with societal relevance.

Subject area

Linguistics or computational linguistics or journalism, media and communication

Specific subject description

The application of techniques from computational linguistics, machine learning, theoretical linguistics, journalism, and/or opinion research for the creation of tools and resources to understand the properties of dog whistles, their political and social effects, their spread, their individual and group behavioral effects, their cognitive implications, all in terms of artificial intelligence, autonomous systems, and agent-based simulation.

Job assignments

This position involves different duties based on the background of the candidate. The winning candidate will enter either the Ph.D. programs in linguistics or computational linguistics at FLoV or the journalism, media, and communication program at JMG. 

We will fund one well-qualified and highly motivated candidate to work on establishing tools (e.g., software), techniques (e.g., algorithms, formal approaches), and/or data resources (e.g., experimental evidence, data collected from media and opinion research) for grounding simulations of the development of dog whistle communication and its effects on individuals (e.g., persons, organizations) and populations (e.g., identity/interest groups, electorates). We plan for this work to be done overall in an agent-based simulation paradigm, where the goal is to iteratively ground highly abstract simulations in increasingly concrete representations of political situations where dog whistle communication may be deployed.

There are many potential “angles of attack” for this work, to be discussed based on the interests and skills of the winning candidate. By way of example, this includes but is not limited to:

• data collection using survey, experiment and interview techniques, particularly in the context of the SOM Institute’s LORE opinion research framework.
• development of the agent-based simulation environment using techniques such as    reinforcement learning and generative adversarial networks.
• use of natural language processing (NLP) techniques to detect patterns in news corpora reflecting patterns identified in a simulation environment.
• integrating game-theoretic approaches into formal representations of dogwhistle use.
   

This work will be done in the context of the GRIPES project and in a highly collaborative environment with other researchers from FLoV, JMG, and allied departments and research groups. It counts as a merit if the candidate as a background or accomplishments in any of the related areas: linguistics (construed broadly; e.g., sociolinguistics, theoretical/formal linguistics), computational linguistics (including machine learning/statistical approaches), and journalism, media, and communication studies. Additional weight will be given to education or work experience in data science, machine learning, natural language processing, agent-based simulation, or advanced statistical analysis methods in any of the GRIPES-related fields (linguistics, computational linguistics, or journalism, media, and communications studies).  

The successful candidate’s main responsibilities will be to write a doctoral dissertation in the areas mentioned above, and to complete obligatory course work. He or she may also undertake a limited amount of teaching, administration, or research not directly connected to his or her dissertation topic. Such duties extend the position accordingly.

The candidate is expected to participate in GRIPES and WASP-HS project activities and courses as well as their home department’s activities. The working language of the GRIPES project is English. There will be funds available for travel as well as computational resources and support for conducting experiments, as relevant. 

Eligibility

To be eligible for the position, the applicant must meet both general and specific entry requirements.

General entry requirements

A person meets the general entry requirements under Chapter 7, Section 39 of the Higher Education Ordinance if he or she:

1. has been awarded a second-cycle qualification, or
2. has satisfied the requirements for courses comprising at least 240 higher education credits of which at least 60 higher education credits were awarded in the second-cycle, or
3. has acquired substantially equivalent knowledge in some other way in Sweden or abroad.

Specific entry requirements

The winning candidate will be admitted into third-cycle programs in either computational linguistics, linguistics, or journalism, media, and communications, depending on qualifications and interests. The admissions criteria for the relevant program will apply.

Admission requirements to the relevant programs are given in the following web links to the general doctoral syllabi:

• Computational linguistics: https://flov.gu.se/digitalAssets/1656/1656982_asp-fd-datalingvistik-2016-eng.pdf 
• Linguistics: https://flov.gu.se/digitalAssets/1685/1685366_asp-fd-lingvistik-2015-eng.pdf
• Journalism, media, and communication studies:
• https://jmg.gu.se/digitalAssets/1651/1651192_fu-eng-syllabus.pdf

Assessment

According to the Higher Education Ordinance Ch.7, applicants will be assessed based on their estimated ability to benefit from the study programs and on the Department’s supervisory resources. We will base this assessment primarily on the following:

1. Theses, degree projects and any publications. These are assessed based on the following quality criteria, of which the final two have the heaviest weighting: Suitable subject definition, relevant questioning, suitable command of the subject and structuring of the thesis, suitable use of language, formalities, academic quality, and independence.
2. A project plan of at most 4 pages excluding references in 11pt font in which the applicant stipulates a research domain that he/she would like to develop, justifies its relevance and discusses which theories, methods and materials would be relevant. The project draft is assessed on the basis of the following criteria: whether the research displays familiarity with previous and current research, whether the research questions are viable and grounded in current research, whether the draft is judged to be realistic, as regards its implementation at third-cycle level and whether the draft shows that it is likely the applicant will make a valuable contribution to research in the field. The project draft also provides supporting information for assessing the Department’s supervisor competence within the research domain stipulated.
3. Other documents such as letters of recommendation.

Admission and selection are also conditional on the Department’s supervisory resources within the doctoral student’s research focus. Admission may also include an interview in addition to a review of qualifications submitted. Admission decisions are made by the Head of Department following preparation at the Department.

Employment

Type of employment: Fixed-term employment

Extent: 100 %

Location: Department of Philosophy, Linguistics and Theory of Science or Department of Journalism, Media, and Communication, Gothenburg

First day of employment: 2020-08-15

For further information please contact:

Asad Sayeed, project leader, asad.sayeed@gu.se

Unions

Union representatives at the University of Gothenburg: http://www.gu.se/english/about_the_university/job-opportunities/union-representatives

How to apply

In order to apply for a position at the University of Gothenburg, you have to register an account in our online recruitment system. It is the responsibility of the applicant to ensure that the application is complete in accordance with the instructions in the job advertisement, and that it is submitted before the deadline. The selection of candidates is made on the basis of the qualifications registered in the application.

Complementary documents, such as publications/books should be sent in two copies to the following address and marked with reference number. University of Gothenburg, Att: HR Administrator, FLoV, Box 200, SE-405 30 Gothenburg, Sweden.

Applications should be written in English. For the application to be considered complete it must include:

1. An introductory letter of 1.000 words maximum (approx. 2 pages), preferably shorter. The letter should consist of four paragraphs in the following order: a) a short presentation of yourself, b) a short presentation of your merits, c) a short presentation of the intended research project and d) a short account for why the GRIPES project is of interest to you.
2. A CV, including a list of publications (if applicable).
3. Theses, degree projects and any publications. These are assessed based on the following five quality criteria, of which the final two have the heaviest weighting: Suitable subject definition, relevant questioning, suitable command of the subject and structuring of the thesis, suitable use of language, formalities, academic quality, and independence.
4. A project plan of at most 4 pages excluding references in 11pt font in which the applicant stipulates a research domain that he/she would like to develop, justifies its relevance and discusses which theories, methods and materials would be relevant. The project draft is assessed on the basis of the following criteria: whether the research displays familiarity with previous and current research, whether the research questions are viable and grounded in current research, whether the draft is judged to be realistic, as regards its implementation at third-cycle level and whether the draft shows that it is likely the applicant will make a valuable contribution to research in the field. The project draft also provides supporting information for assessing the Department’s supervisor competence within the research domain stipulated.
5. A short list of 2-3 academic or employment references (optional).

Closing date: 2020-04-15

The University of Gothenburg promotes equal opportunities, equality and diversity.

Applications will be destroyed or returned (upon request) two years after the decision of employment has become final.