What risk is there that generative AI evolves such that the content it generates can avoid detection faster than tools can be developed to detect it? How can international and industry collaboration limit this risk?
Background
Although there is already material evidence on the types of serious harms individuals encounter online, there still remain a number of emerging harms, where the evidence base is still yet to mature (e.g. epilepsy trolling, online animal abuse). SOH would like to close this significant gap in understanding the impact of encountering different types of serious harms online and understanding the best approaches to measuring the impact of the Online Safety legislation.
SOH highlights the importance of Media Literacy in the digital age and asks for further studies to uncover barriers to engagement as well as the effectiveness of DSIT programmes. This issue closely relates to Counter-Disinformation interventions, which requires evidence for its effect on bystanders, topic specific disinformation and what tools can be used to combat this issue.
Research on Safety Technology would greatly develop SOH’s understanding of the relationship that DSIT online safety objectives have with the technology market today. A primary focus lands on improving Age Assurance (AA) measures. This includes ensuring transparency and assessing opportunities for the sector.
Next steps
If you are keen to register your interest in working and connecting with DSIT Digital Technology and Telecoms Group and/or submitting evidence, then please complete the DSIT-ARI Evidence survey - https://dsit.qualtrics.com/jfe/form/SV_cDfmK2OukVAnirs.
Please view full details: https://www.gov.uk/government/publications/department-for-science-innovation-and-technology-areas-of-research-interest/dsit-areas-of-research-interest-2024
Source
This question was published as part of the set of ARIs in this document:
Topics
Related UKRI funded projects
-
AI Safety Platform: Generative AI and Cybersecurity Training SaaS for Schools and Families
**Problem statement:** The rapid rise of **generative AI** has introduced **unprecedented cybersecurity risks**, particularly for students and families. **Deepfakes, AI-driven scams, misinformation, identity theft, and c...
Funded by: Innovate UK
Lead research organisation: UNIVERSITY OF EAST LONDON
Why might this be relevant?
The project specifically addresses the risks of generative AI and provides solutions for detection and prevention.
-
Integrating user experience data into image algorithms to mitigate online harm
The culmination of decades of academic research and commercial application, this proposal offers a step change on how algorithms account for the end user experience. Images hold within them varying degrees of emotional '...
Funded by: Innovate UK
Lead research organisation: RHIZOMETRIC DESIGN LIMITED
Why might this be relevant?
The project aims to mitigate online harm through image algorithms, which is related to online safety but does not directly address the risks of generative AI.
-
RAIX: An add on to make the content creators application safely and trustworthy through Responsible AI toolkit
The UK has a robust digital content creation industry encompassing various sectors such as social media, blogging, video production, podcasting, and online publishing. This industry continues to grow as more individuals ...
Funded by: Innovate UK
Lead research organisation: SAFE TECHNOLOGIA LIMITED
Why might this be relevant?
Partially relevant as it focuses on content creators online safety, but does not specifically address generative AI evolving faster than detection tools.