PhotoVoice has published a new methodology paper, titled Participatory Photography, Photography, and AI.
This PhotoVoice Methodology paper attempts to address some of the concerns and questions raised by practitioners, academics, and community members about the use of AI, as well as develop a set of principles around these considerations.
It is intended to provide an accessible entry point into how AI works, how photography has been impacted by AI, and the use of AI when delivering community engagement, research, and participatory practices.
It includes sections considering:
- What is AI and why is it relevant?
- AI and photographic processes, including generative AI and computational photography
- Image manipulation
- The use of AI in analysis and research.
The paper concludes with a set of principles which outline PhotoVoice’s policy on the use of AI in its projects and activities.
The full paper is available to PhotoVoice Supporters, with the principles shared below. These principles will be incorporated into future versions of the PhotoVoice Statement of Ethical Practice.
PhotoVoice Principles on the Use of AI in Participatory Photography Projects
- We recognise that the subjective nature of photography, the individual perspective of the photographer, and the availability and use of different forms of technology will mean that the same subject at the same time can be photographed in very different ways.
- We believe that the photographer’s intention should be the primary point of insight into understanding what these differences are and why they have occurred.
- We do not support the specific manipulation of a photograph – including its accompanying contextual information – which substantially alters, amends or misrepresents what was taken by the photographer at the point of creation.
- We recognise that specific choices may be made by the photographer to alter how the resulting image appears, which may include choices around framing, cropping, colour, and contextualisation. This should only be done to enhance the photographer’s intention, not to mislead or misrepresent what was being taken.
- We recognise that the taking of photographs will be increasingly reliant on some degree of manipulation undertaken at the point of creation, due to developments in technology and availability of equipment, specifically around Computational Photography.
- AI should not be used to analyse, infer, or draw conclusions about why a photograph was taken. Analysis of a photograph should only be undertaken with the photographer themselves. Additional analysis should not be undertaken which seeks to provide additional insight or conclusions outside of this.
- We do not advocate for the generation of content using AI unless it is a) undertaken for a clear and specific purpose which could not be reasonably achieved in any other way and b) clearly contextualised as having been done so. We believe that the creation of such content should be clearly identified and labelled. We do not believe that wholly or substantially AI-generated content should ever be passed off as original.
- We will not use AI to generate accompanying material around our activities or projects, unless there is a clear and compelling reason to do so. If it is undertaken, we will be transparent that this has been the case and explain the reasons why.
- We recognise photography as a powerful form of self-expression which has the potential to provoke strong reactions in those who see it. We should never seek to mislead. However, influencing and encouraging people to act in response to participant photographs will often form a part of our activities.
- Those viewing the outputs and content of our projects and activities should feel confident that what they are seeing was created by humans.
These principles can be downloaded as a free PDF here. If you’re a PhotoVoice Supporter, you can access the full paper for free here.
