Posted: 13/08/19
ADAPT members will play an active role in the upcoming WorldCon. Taking place in the Convention Centre from the 15th – 19th of August, World Science Fiction Convention is the longest-running Science Fiction convention in the world.
Professor Dave Lewis, Associate Director in ADAPT will present ‘No Datafication Without Representation!’, a paper on the power of modern data-driven AI technology. Academics, companies, governments, and civil society are becoming increasingly alert to the need for trustworthy AI. Trustworthy AI must deliver AI bias, respect for privacy and human autonomy, robustness and safety, transparency and accountability. The balance between technological solutions, self-regulation, and government regulation in achieving trustworthy AI will be the defining debate in the field in the coming years.
Dr Maire Brophy, Strategy Lead for Research Funding within ADAPT will chair a panel on ‘Does an AI need a body?’. Joined on the panel by Dave Lewis along with Maquel A. Jacob and Manny Frishberg, they will discuss the topic of if an AI needs a body, or do we need to see an AI in a body to recognise it is ‘alive’ or sentient? Why are we so obsessed with assigning gender, or gender characteristics, to an entity that is basically a software spark in a hardware system that doesn’t rely upon organic reproductive systems?
Dr Kevin Kiodl will host an exciting workshop entitled ‘Would you swap your partner with a Robot designed to be ‘the one’?’ Since the beginning of society, we strive for relationships with others. Science tells us that we do this to increase our chances of survival. Yet, even in steady relationships, many of us feel misunderstood, unwanted and ultimately not loved. Why is that? What if there is that perfect partner out there somewhere? What if you could build ‘the one’? Would you replace your partner? Join the Heliosphere for an all-inclusive debate about the future of meaningful relationships and how technology will change it. The Heliosphere revolves around a 360-degree camera which allows online audiences to partake in the debate. The physically present audience sits in a circle around the main round table at which a limited number of members debate directly. The inner-circle can vacant their place at any time to let members of the outer circle join. To gain a real-time overview state of the art Speech-to-text extraction summarizes the topics discussed.
If you are attending WorldCon 2019, be sure to check out the multiple panels and talks with ADAPT members. Key events listed below.
Dr Maire Brophy
Introduction to Irish SFF 15 Aug 2019, Thursday 14:00 – 14:50, Wicklow Hall-1 (CCD)
This panel is the ideal starting place for anyone unfamiliar with Irish SFF. There will be a brief overview of the genre as a whole and how it relates to Irish culture and history, as well as reading recommendations. Later in the convention, there will be panels covering the topic in more detail from a variety of viewpoints. (Panel: Nick Larter (M), Dr Máire Brophy (ADAPT Centre), Dr Jack Fennell, Ruth Frances Long)
Does an AI need a body? 15 Aug 2019, Thursday 20:00 – 20:50, Liffey Hall-2 (CCD)
Does an AI need a body, or do we need to see an AI in a body to recognise it is ‘alive’ or sentient? Why are we so obsessed with assigning gender, or gender characteristics, to an entity that is basically a software spark in a hardware system that doesn’t rely upon organic reproductive systems? (Panel: Prof Dave Lewis (ADAPT Centre, School of Computer Science and Statistics, Trinity College Dublin) Maquel A. Jacob (MAJart Works), Manny Frishberg (Wordfire Press, Sky Warrior Books, Knotted Road Press, B Cubed Press), Dr Máire Brophy (ADAPT Centre) (M))
Autographs: Monday at 10:00 19 Aug 2019, Monday 10:00 – 10:50, Level 4 Foyer (CCD) G. David Nordley (The Experimental Rocket Propulsion Society, USAF (ret.)), Laurel Anne Hill, Dr Máire Brophy (ADAPT Centre), Jenn Lyons, Stefan Rudnicki (Skyboat Media Audiobooks), Alexander Dan Vilhjálmsson (Gollancz)
Well-meaning vs ‘plain dealing’ villains 19 Aug 2019, Monday 12:00 – 12:50, Wicklow Room-1 (CCD)
Villains come in many different guises. Characters like Marvel’s Thanos do terrible things for what they believe to be good reasons, whereas The Other Mother from Coraline is alien and terrifying. Do certain genres lend themselves to different types of villainy? Are villainous women held to a different standard than their male counterparts? (Panel: Russell Blackford (University of Newcastle, NSW) (M), Dr Máire Brophy (ADAPT Centre), Lauren Roy, Paul Anthony Shortt)
Prof Dave Lewis
‘No Datafication Without Representation!’ 18 Aug 2019, Sunday 10:30 – 11:20, Odeon 6 (Academic) (Point Square Dublin) As awareness of the power of modern data-driven AI technology grows, academics, companies, governments, and civil society are becoming increasingly alert to the need for trustworthy AI. Trustworthy AI must deliver AI bias, respect for privacy and human autonomy, robustness and safety, transparency and accountability. The balance between technological solutions, self-regulation, and government regulation in achieving trustworthy AI will be the defining debate in the field in the coming years.
Regulation and rules governing AI are a well-established theme in science fiction, from Asimov’s three laws to Gibson’s Turing Police. However, several countervailing real-world factors may make trustworthy AI through regulation unlikely. These include:
the spread of technical skills and computing power needed to build AI easy access to AI services across regulatory jurisdictions the asymmetry in power between large organisations using AI and the people they impact the opaque nature of modern deep learning AI the deep technical and legal expertise that large companies can bring to bear on any regulatory conflicts political aversion to the perceived chilling effect of regulation of innovation and competitiveness the slow pace of regulatory development relative to technological innovation.
An alternative for exerting societal control over AI is through expanding the rights of people to own data. Data is the fuel on which the effectiveness of AI is based, and the impact of any AI application grows in step with the relevance and volume of data available to train the AI. The stakeholders sourcing data for an AI system are often the most impacted by that AI system. Enforcing trustworthy AI measures before the AI can leverage sources of data will reflect stakeholder needs more accurately, and scale with the number of stakeholders more naturally, than doing so after the data is acquired and the AI starts impacting people’s lives.
There is currently no universal legal basis for data ownership, but instead, a complex set of rights based on whether, for example, the data identifies a person, represents an original creative act or is a defensible form of intellectual property.
One proposed universal data ownership right suggests the non-exclusive ownership of any data for any party that can show a traceability log of having produced or contributed to the data. This enables individuals, families, patient groups, organisations, cities, countries, etc. to negotiate collectively on the implementation of trustworthy AI in return for permission to use their data. Therefore the agreement on implementing trustworthy AI requirements works with the flow of data driving AI innovation, rather than relying on an external regulatory system to keep up with it. This simple universal data ownership measure, therefore, offers opportunities for new forms of data ownership-based polities to be imagined in the struggle to ensure there is no datafication without representation. part of: AI
1. Prof Dave Lewis – ‘No Datafication Without Representation!’ 2. Prof David Powers – ‘AI, Talking Robots/Ships, Brain Control Interface’ 3. Jo Lindsay Walton – ‘Project Cybersyn’
VR and AR: the future of gaming or a fad? 15 Aug 2019, Thursday 12:00 – 12:50, Wicklow Room-1 (CCD)
In 2015 the gaming industry was abuzz with the potential of both virtual and augmented reality. However, the expected explosive growth of these technologies did not happen. This panel will discuss the potential of these exciting technologies and whether or not they will ever achieve the potential people believed they would have a few short years ago. (Panel: Mr Kevin Elliott (Oxford University), Rebecca Gomez Farrell , Prof Dave Lewis (ADAPT Centre, School of Computer Science and Statistics, Trinity College Dublin), Tom Toner (Gollancz) (M))
Censorship in SFF and Comics 16 Aug 2019, Friday 17:00 – 17:50, Liffey Room-2 (CCD)
Our genre has long been subject to censorship and control (D&D, monster movies). This has led to suppression or trends towards “vanilla” (safe) content. How do we protect creative expression? How much should we hold to account companies that profit from characters dedicated to justice, inclusion, and diversity while pandering to consumers who are against all of that? (Panel: Roz Kaveney, Prof Dave Lewis (ADAPT Centre, School of Computer Science and Statistics, Trinity College Dublin) (M), Sienna Saint-Cyr (SinCyr Publishing L.L.C.), Mikko Rauhala.)
No, what do you mean by AI? 19 Aug 2019, Monday 10:00 – 10:50, Liffey Hall-2 (CCD)
AI professionals generally consider HAL-like general artificial intelligence (the AI that worries Elon Musk) so far in the future and so poorly understood that it’s useful only for philosophy and entertainment. Intelligence as we know it may be irrelevant. Neither is AI necessarily automated: there’s lots of human validation in AI, in supplying training data and validating results. Panellists discuss how current and near-future states of AI differ from their fictional counterparts. (Panel: Renee Sieber (McGill University), Dr Tamara Vardomskaya (M), Prof Dave Lewis (ADAPT Centre, School of Computer Science and Statistics, Trinity College Dublin), Kevin Roche (IBM Research, Worldcon 76).)
Share this article:
