COGS200GRP26
Group ideas
1. Ambient intelligence:
Ambient intelligence is pervasive in contemporary society, with a diverse accessibility to this form of intelligence. How people interface with ambient intelligence, and how it changes our ideas of intelligence. As well, monitor how human behavior is influenced by the presence of ambient intelligence. Fragmentation of individual intelligence and outsourcing to computers for storage, or the extension of human memory in technology.
2. Development / Raised by Robots
Addresses questions of nature vs nurture (PSYC and PHIL aspects): Are we predetermined to be a certain way? Can this be tied with ambient intelligence? Are parental responsibilities being outsourced to technologies?
3. Artificial Intelligence learning through site-specific channels
Address the difference in languages expressed, behaviour of computation, or any derived motivation.
4. Exploring potential of physical biomarkers of consciousness,
and applications of unconscious patterns to diagnoses of psychological disorders such as schizophrenia
5. Facial recognition of emotion
3D facial recognition to trace and recognizes patterns of emotions, and develop strategies to prevent dangerous behavior such as suicide, terrorism, violence, etc... Must define behavioral patterns
6. Instant language translational software
Applied usages of the software such as note-taking, or translating speech to text.
7. Modeling emotional aspect in technology
Train computer modules to react or behave in certain ways, or being able to give technology some motivational drive.
8. Deriving language from non-human communication such as dogs, cats, etc...
9. Implementing electrodes or brain implants, and determining if neural patterns exist among disorders. As an extension, determine if there is a potential remedy to any disorders
10. Perfect Song Program
Effect of repetition (Linguistics) on neural response. Developing a personal curated music app, which monitors your behavior in relation to musical reception. Learning from behavioral patterns, the device will predict personal moods and seek for appropriate musical tracks.
Possible combined directions
Effect of AI on human cognition: Does frequent interface with computers change the way our minds/brains work? How are children who are 'raised by robots' different compared to children who aren't? Could studying this provide us with means of locating biomarkers of consciousness in the brain? I.e.; is consciousness necessarily tied to cognition? If so, which aspects? If using technology to outsource aspects of our cognition affects the ways our minds or brains function, how does this alter our understanding of intelligence?
AI and emotion: Can a computer have emotions? Could one be trained to 'feel' certain ways and to react accordingly? Could this make machines better able to recognize emotional markers in people (e.g. facial expressions), and predict their behaviours? What are the possible applications? What could a machine's capacity for emotion tell us anything about human emotion?
AI and language: How can AI's current abilities to 'understand' language be used to improve our lives? What other 'languages' can a machine be taught to understand (music, animal sounds), and could they be translated to languages we can understand? What are the differences in the ways AI process different languages?
Predictive ability of AI: Can computers reliably predict people's behaviours? How can this be applied? Facial expressions as predictors of behaviour, using past behaviour to predict note-taking or musical preferences across a range of contexts, etc. If our behaviours are able to be modeled and reliably predicted by a computer, what does that contribute to our understanding of human behaviour and cognition of it?
Possible details we can discuss regarding to combined directions
Effect of AI on human cognition:
- How are children who are 'raised by robots' different compared to children who aren't? - Are the children "raised by robots" different from others in developing a close relationship in their adulthood? Based on attachment theory, we know that early childhood experience, specifically interacting with their caregivers (usually parents, here in our case may be the robots) has longterm impacts on their future attachment behaviors.
AI and emotion:
- emotion recognition in specific context - possible application: criminal investigation, psychologic diagnosis..
Finalized Project Pitch
Ambient Intelligence (AmI) refers to information and intelligence embedded within hidden networks of interconnected systems. These systems are widely available (e.g., Smart Home Monitors, FitBit, Cloud computing, etc…), and are characterized by technologies that are ubiquitous, transparent, and intelligent. Moreover, current AmI developments aim to exponentiate its ubiquity, and provide infrastructure supporting universal access points.
However, research regarding societal impacts of AmI is still in its infancy, which is further compounded by rapid advancements in technology. As well, current studies of the impact of AmI on cognitive development are largely piecewise. In this study, we implement a longitudinal cohort study to monitor participant households employing standardized levels of AmI. The comparative study targets the cognitive development of children in the presence and absence of AmI. We focus on child development to capitalize on features such as critical periods, impressionability, and novelty of AmI exposure.
As this research project aims to elucidate effects on cognition correlated with AmI, it could provide invaluable evidence for the development of philosophical frameworks concerned with the interaction between humans and intelligent machines. For example, assume that the study finds that children raised in AmI environments show marked differences in cognitive performance than a control sample when both and when neither group has access to the same technology. Data of this type could support theories of cognition like the Extended Mind Hypothesis and allow deeper investigations of its implications for individuals who act as part of a cognitive system that extends beyond their physical body, or it could indicate the inadequacy of such theories and drive the development of more accurate frameworks. Further, the study may also produce evidence which would support further investigations of ethical considerations important to the implementation of AmI algorithms that do not impose on human freedoms, especially in weighing consumer interests against those of corporate and commercial entities.
Child Development in Environments Saturated by Ambient Intelligence
Greenfield, P. M. & Calvert, S. L. Electronic media and human development: The legacy of Rodney R. Cocking. Journal of Applied Developmental Psychology 25, 627–631 (2004).
Media consumption and interaction is predominant among developing children. Often, children engage in multiple media-intensive activities simultaneously. The paper addresses the effect of AmI on children’s cognitive, emotional, and social development. The summary focuses on interactive media’s, which is a hallmark of newer technologies, effect on children’s “representational competence” (iconic representation, mental rotation, spatial visualization), learning capacities, and formation of identities. Evidence supports skill with and use of iconic representations, such as symbols in AmI applications, facilitate scientific-technical learning. That is, AmI provides an informal learning environment that primes children’s representational skills of the world’s science and technology. Evidence also suggests that interactive media bridges the gap between representational acts and actual, physical behaviour. That is, engaging in activities via AmI is more powerful than viewing-only since it translates easily into a child’s behaviours.
Ma, W., Adesope, O.O., Nesbit, J.C., & Liu, Q. Intelligent tutoring systems and learning outcomes: A meta-analysis. Journal of Educational Psychology, 106, 901-918 (2014).
Intelligent Tutor Systems model individual psychological states to tailor individualized instruction. ITS-learners performed significantly better than those who learned via 1) large-group, teacher-led instruction; 2) non-ITS computer-based instruction; 3) and textbook/workbook instruction. However, no significant difference was found when compared with individualized human tutoring or small-group teacher instruction. In relation to our project, can an “Intelligent Parenting System” be a substitute for human parenting? Would such a system negatively affect a child’s cognitive functionality and development? Or would such a system show negligible effects on a child's performance, as in the ITSs studied in the paper?
Trepte S., Reinecke L. Privacy Online: Perspectives on Privacy and Self-Disclosure in the Social Web. Springer Science & Business Media (2011).
Investigates the nature and ethics of online privacy and how users are willing to abandon privacy in favour of social gratification. Shows how people of different ages deal with privacy using a developmental and sociological perspective. In relation to our project, we can use the data they presented to consider how children raised in AmbI are willing to disclose information to non-human child-raisers. This book also deals with philosophical problems like the appropriate extent of privacy and the privacy paradox.
Investigative Approaches for Studying Children Raised Under the Environment of Ambient Intelligence
Kable J.W., Caulfield M.K., Falcone M., & et al. No Effect of Commercial Cognitive Training on Brain Activity, Choice Behavior, or Cognitive Performance. J. Neurosci. 37, 7390-7402 (2017).
Commercial cognitive training apps, such as Lumosity, claim to improve mental processing. A cognitive assessment and fMRI analysis of adult performance during decision-making tasks, including delay discounting (small immediate rewards vs. large delayed rewards) and risk sensitivity (small guaranteed rewards vs. large risky rewards) tasks. Research suggests cognitive training has no effect on neural activity in decision-making beyond those tasks specifically trained; that is, the “cognitive training” is not transferrable to other mental processes. In relation to our project, we could apply the methodology to children performance during decision-making tasks. However, it’s arguable whether children have developed cognitively enough to differentiate between reward types, which makes it difficult to extract useful information. Controls: Random participants played online sources of non-cognitive training games (games that do not test executive function and are non-adaptive); Performance assessed pre-training, and post-training Independent: Random participants played cognitive-training games (Lumosity) Dependent: Performance in cognitive tasks focusing on attention (continuous performance task), memory (visual/spatial n-back), response-inhibition (stop signal task), interference control (Stroop test), and cognitive flexibility (color shape test).
Farrell M.E., Kennedy K.M., Rodrigue K.M., & et al. Association of longitudinal cognitive decline with amyloid burden in middle-aged and older adults: Evidence for a dose-response relationship. JAMA Neurology 74, 830–838 (2017).
Research analyzed a correlation between cognitive decline (tasks targeting episodic memory, reasoning, processing speed, and vocabulary) and amyloid burden. The research provides a method (in vivo amyloid imaging via PET scans) to monitor cognitive performance. In relation to our project, we may instantiate a study to monitor amyloid development in children interfacing with AmI (Ambient Intelligence) to determine cognitive effects of AmI. Alternatively, we may monitor other biomarkers using PET imaging techniques.
Barbuscia, A., and Mills, M.C. Cognitive development in children up to age 11 years born after ART—a longitudinal cohort study. Hum Reprod 32, 1482–1488 (2017).
Study compares cognitive capacity of children born by natural conception or by assisted reproduction technology (ART), such as IVF and ICSI. Research team focuses on latent growth curve models and considers confounding parental characteristics in their study. Cognitive development is assessed by “British Ability Scales”, which comprises of 1) naming vocabulary task to measure expressive verbal ability (ages 3-5); 2) word reading test to measure verbal cognitive abilities (ages 7-10); 3) verbal similarity test to measure verbal cognitive abilities (ages 11). Confounding variables considered include parental health, age, and socioeconomic status. In relation to our project, we may adapt the “British Ability Scales” as a method to assess cognitive development. The paper also directs attention to confounders or limitations of our project: 1) parental characteristics of children raised by, or not raised by, AmI; 2) Accessibility to a large candidate population raised by AmI; 3) Difference in sample size between groups raised by AmI and groups not raised by AmI.
Dresler, M., Shirer, W.R., Konrad, B.N., & et al. Mnemonic Training Reshapes Brain Networks to Support Superior Memory. Neuron 93, 1227–1235.e6 (2017).
Paper examines the functional brain network organization of participants via fMRI scans. Research found functional connectivity changes after induced training. In relation to our project, we could examine the brain’s functional network reorganization in children pre- and post-exposure to AmI. This may address concerns of whether long-term contact with AmI reduces functional connectivity, or produces a brain network organization that is not conducive to desirable cognitive performance. For instance, does extending personal memory via artificial agents, and/or relying on access to AmI for solutions affect the brain’s ability effectively reorganize neural activity?
Druga S., Williams R. “Hey Google, is it OK if I eat you?”: Initial Explorations in Child-Agent Interaction. IDC Stanford (2017)
Research on how children view intelligent machines. Dependent variables include how well children trust agents like Google Home, Julie Chatbot, Cozmo, and Amazon Alexa. Also presents data on how children play with and understand these agents. In relation to our project, we can use this research to see the extent of an intelligent system’s ability to raise children. This is possible by seeing what parental aspects these agents can express, including their ability to answer children’s questions, how much children trust these agents, and how intelligent children view them.
Significance of the Proposed Research Project
Heckman, J.J. Skill Formation and the Economics of Investing in Disadvantaged Children. Science 312, 1900–1902 (2006).
A summary of early environmental influences on children cognition. Presence of family environments support development of cognitive and non-cognitive abilities (motivation, perseverance, tenacity, punctuality, etc…). Disadvantaged environments are predictors for adult failure, and results in social and economic problems such as illiteracy and criminal activity. In relation to our project, studying children raised by AmI may provide useful insight on the environmental status supported by AmI. In other words, does AmI create a “disadvantaged environment”? Does AmI support development of cognitive and non-cognitive abilities? Alternatively, does AmI create an imbalance between cognitive and non-cognitive abilities? Addressing these questions may have practical ramifications on policies and programs directed at early childhood intervention.
Vulchanova, M., Baggio, G., Cangelosi, A. & Smith, L. Editorial: Language Development in the Digital Age. Front Hum Neurosci 11, (2017).
Child-robot interactions show children are able to differentiate among robotic informants, and interaction relies on robotic features (coordinated responses, adaptive behaviour, effective feedback). Currently, children-robot interaction without parental mediation is poorly studied. There is a need for a holistic approach to cognitive development, especially in language development. However, such a model is complex and requires dense collections of data. Our project would supplement a longitudinal data collection and analysis of the affected development of those (children) whom interface with AmI. For instance, whether the effects of digital tools, in this context, are temporary or long-lasting. This type of research may also inform future AmI designs, such as in assistive robotics purposed towards helping children with cognitive or motor deficits.
Children raised under the environment of ambient intelligence
'A short draft of introduction：'' Why - Justification for our research
With prevalence of ambient intelligence, children are being exposed to AMI environment at an increasingly early age. They may interact with AMI across different circumstances since their births. For preschool age children, they spend the majority of time at home where they may start to utilize some home appliances which are often embedded with computing power (e.g., washing machines, refrigerators, TVs, air conditions) (Cook, 2009) . Moreover, as the AMI plays a significant role in the interactive games and toys designed for children, it may have meaningful impacts on children’s abilities, skills, and other developments over time (Ioannidi,2017). For school-age children, interactions with AMI continue to grow. Lots of schools are adapting to more advanced education modes, and the embedding of teaching-learning process changes the traditional teaching-learning process. These shifts could change children’s learning abilities, and behaviours, too.
Questions that we can address / What do we want to measure?
Does interacting with AMI affect children‘s cognitive ability?
eg : Impacts on learning ability :
Does interacting with AMI affect children‘s interpersonal ability/skills?
Research / Experimental design
Cohort design, cross sectional design?
Ambient intelligence: Technologies, applications, and opportunities
Security-enhanced ambient assisted living supporting school activities during hospitalisation
http://www.sciencedirect.com/science/article/pii/S157411920900025X
Designing Games for Children with developmental disabilities in Ambient Intelligence Environments
http://www.sciencedirect.com/science/article/pii/S2212868916300290
More miscellaneous sources
Effects of internet use on the adolescent brain
http://web.b.ebscohost.com.ezproxy.library.ubc.ca/ehost/detail/detail?vid=0&sid=e73d0e4a-f752-4f2a-920e-69851a35ec96%40sessionmgr104&bdata=JnNpdGU9ZWhvc3QtbGl2ZSZzY29wZT1zaXRl#AN=2014-31646-002&db=psyh Not enough evidence to say that internet use in adolescents significantly affects brain function, however studies suggest that this time is one for social development. Studies have not shown internet use to impair socialization, but still a concern? Very nuanced. Some effects on cognition have been hinted at, but need more attention.
Google effects on memory
http://science.sciencemag.org/content/333/6043/776.full People are becoming more disposed to remembering where information is available rather than the information itself as long as they have the knowledge that they will be able to access it later. Suggests that Google (and search engines generally) have become a form of transactive memory: an external, online extension of our own memory.
Analytical reasoning, learning in social networks
http://rsif.royalsocietypublishing.org/content/royinterface/11/93/20131211.full.pdf Studied learning using connected computer terminals (kind of simulates internet). Study suggests that answers to specific questions can be learned through networks, but the analytical skills to arrive at the answer independently (contagion of analytical processing) cannot be learned through exposure to the answers alone (even though analytical approaches can be primed for with very subtle cues). The internet may allow us to better solve analytical problems collaboratively, but reduces the frequency of individual analytical contributions. However, these are results of copying and not teaching.
Seven fears of mobile technology affecting adolescents
Why the biggest challenge facing AI is an ethical one
A brief overview of ethical problems raised by the rise of AI. Can be a good jumping off point to start research.
http://www.bbc.com/future/story/20170307-the-ethical-challenge-facing-artificial-intelligence
Kids should not be guinea pigs: Mattel pulls AI babysitter
A company forced to cancel development of an AI that could babysit by responding to crying babies, aid in developing language, and promote good behaviour.
https://www.theguardian.com/technology/2017/oct/06/mattel-aristotle-ai-babysitter-children-campaign
Kids, AI devices, and intelligent toys
Very useful research article on how machines affect child development.
https://www.media.mit.edu/posts/kids-ai-devices/
What’s next for research on young children’s interactive media? http://web.b.ebscohost.com.ezproxy.library.ubc.ca/ehost/detail/detail?vid=0&sid=45a4e1c6-102d-4c52-889c-aeab9bac17dc%40sessionmgr120&bdata=JnNpdGU9ZWhvc3QtbGl2ZSZzY29wZT1zaXRl#AN=2016-03094-008&db=psyh Good general review of research on media aimed at children. Offers a particularly hopeful view of interactive technology as a valuable learning tool. Highlights the need for more research in the area, particularly because of the accelerated speed of technological developments.
Cognitoys https://cognitoys.com/ Learning companion toy for children. Smart toy that adapts to children’s question-asking behaviour.
Impact of video games on hippocampus https://www.nature.com/mp/journal/vaop/ncurrent/full/mp2017155a.html Found reduced grey matter in subjects playing action video games. Suggest that this is because the games use gps and waypoints to direct the player, so that the cingulate gyrus is activated during navigation more than the hippocampus, which would normally be acting during navigation. (https://search-proquest-com.ezproxy.library.ubc.ca/docview/1933223936?pq-origsite=summon&accountid=14656 for more info)