Social BRIDGES: The near-future of AI: How will humans and AI interact in 5 years?

In the upcoming socialBRIDGES e-conference on 21-23 April 2021, we are pleased to offer an interdisciplinary three day event on human-AI interaction consisting of 20 minute talks, 45 minute keynote talks, and a virtual poster room.

The event features

  • keynote speeches by leading experts in the field
  • short presentations
  • a virtual poster room

 

New to socialBRIDGES, there will be a GatherTown virtual space for participants and guests to socially interact during coffee breaks, as we would do during real events.

The event will focus on three key themes.

 

Theme 1

AI beyond tools: Cooperating and competing with artificial non-human agents

When computer scientists develop a machine-learning application or even an AI system, they do so to solve a specific problem: driving a car, recognising faces, or finding game-winning strategies. In these circumstances AI systems are nothing more than tools. We, the users of machines, have goals and machines are there to help us meet them. But when it comes to human-AI cooperative contexts, should we treat AI as more than tools and when is it beneficial or even necessary to include a ‘human-in-the-loop’? 

Examples of research questions:

- How does the make-up of human and non-human groups change how we coordinate our actions?

- Will we trust, expose ourselves to risk, and cooperate with artificial agents as much as we do with fellow humans?

- What challenges do we face in ensuring that the introduction of AI systems into our society is as smooth and efficient as possible?

- Are there obstacles to our cooperation with machines and are there ways to overcome them?

- How to foster our trust in AI systems? Are there reasons not to trust them? Can these worries be overcome and how?

- We want AI systems and their use to be unconditionally benevolent, explainable, and fully transparent. Is that achievable? Are there contexts in which we would not want that?

 

Theme 2

Extending the senses: Learning and perceiving in human and non-human agents

Given adequate training data and time to learn, current machine learning applications can rival human performance. However, while the performance outputs like in visual object recognition are comparable, the underlying sensory processing is not. Machine learning performance is stifled when objects are rotated or some pixels are altered, whereas human perception is vulnerable to optical illusions. Here, we discuss the similarities and differences in learning and perceiving in human and non-human agents.

Examples of research questions:

- Can we couple artificial sensors and computation with human perception?

- Should we think of, and can we design AI perception to be like human perception?

- Can human perception and cognition be enhanced with the use of smart technologies?

- Can we use AI technologies to apply our senses in contexts in which we have not used them before?

 

Theme 3

Robots like us: Machines that look and think like humans

While this could be the sequel to Ian McEwan’s ‘Machines like me’, here we want to turn our attention to issues of embodied cognition: Are these robots capable of social interaction?  Possible application areas of robotic AI systems involve care-homes, the military and the modern workforce. But should we use them for this purpose and, if so, what should we be thinking about as we do so?

Examples of research questions:

- Do robots have to think and look like us for us to be able to coordinate actions effectively?

- Why and when do we (want to) anthropomorphize machines?

- Is it enough for us to believe that machines are similar to us in order for us to trust and welcome their use, or do we have to know that they are in fact similar to us too?

- In human-machine interaction, can we nudge people into believing that they interact with someone “like” them? Are there contexts in which we should or should not do that?

 

We welcome one and all to submit abstracts for talks or our virtual poster wall. Videos of both talks and posters will be made available after the conference.

The event activities will be live streamed on the day via YouTube. Depending on participant consent, videos of talks and posters will be made available after the conference on our Social BRIDGES YouTube channel.

The online event is hosted by the Universität der Bundeswehr München.

To get a taste of our previous Social BRIDGES events, click here

 

Stay tuned for more information and programme updates!

 

Our communication rules

BRIDGES is committed to facilitating the exchange between established and early career researchers. Additionally, to counter a propensity to propagate at lightning speed the news we want to hear, we believe in clear, precise and accessible scientific communication to ensure that the results of our community’s output are communicated so as best to serve our public funders. To that end, a highlight of our e-conferences will be a competition, sponsored by Gorilla.

Our communication rules

About SocialBRIDGES

SocialBRIDGES is the first in a series of e-Conferences hosted by the University of the Bundeswehr, Munich and actively supported by the following individuals:

Organising committee

Prof. Merle Fairhurst, Universität der Bundeswehr München
Prof. Ophelia Deroy, LMU Munich
Dr. Jurgis Karpus, LMU Munich
Dr. Maximilian Moll, Universität der Bundeswehr München
Louis Longin, LMU Munich
Dr. Olga Lantukhova, Universität der Bundeswehr München

Supported by following institutions:

Supporters Social BRIDGES 3.png