OUR SCIENTIFIC BACKGROUND
[ picture from the LTNS team archives ]
Our work connects social research, psychology and responsible technology development. This background informs the questions we ask; it does not establish the effectiveness of the planned application.
We develop LTNS as a collective effort across the world, with different cultural perspectives helping shape the design and interpretation of our research.
OUR INNOVATION: VIDEO, VOICE AND HUMAN SUPPORT
LTNS doesn't bring experience for adults who did not get to say goodbye to someone absent from their life. Our product concept (Year One) is: still from an owned solo file - labelled generated clip - stream only. Research protocol: some studies may add in the future a professional before/after. That protocol is not public app.
Our first product concept is a six-second AI-generated clip saying "Long Time No See...". We start with source video containing audible speech, so the design can draw on expression, movement and voice together.
START SMALL. EVALUATE CAREFULLY.
The first stage focuses on one short, clearly labelled experience. Longer user-written videos and controlled real-time conversations belong to later stages of our roadmap. The application is in development.
PREPARE, EXPERIENCE, REFLECT
Experience only if appropriate: review the generated material, choose whether to watch and keep an immediate stop option. AI cannot speak for the absent person.
Reflect with support: allow time to discuss the experience and decide what comes next, including choosing no further AI use.
INCLUSIVE RESEARCH ACROSS THE WORLD
Our team aims to involve different cultural perspectives and lived experiences in design, evaluation and interpretation. We want to assess emotional responses, acceptability, accessibility and possible harms, compare alternatives and follow outcomes over time.
Reviews and transparent reporting are part of the proposed research process. Innovation should remain accountable to the people it is intended to serve.
OUR STARTING POINT: NAIROBI, 2023
The presentation by Grazyna Czubinska and Roman Mazur related to psychological perspectives in applying safe and ethical AI. LTNS concept is shaped by a wider international team.
To be brief in progress:
The six gates. “Reject” cases (second face, child, celebrity, unowned file).
Clone / resurrection / last words / therapy / downloadable deepfake / public share / grief ads.
Age, ownership warranty, stream-only, watermark, delete, no training, family objection process, crisis signposts UK then US.
UK GDPR for the living user. Absent person is not a UK data subject (the user’s file and account are). Retention periods. UK/EU region. US users: we will publish a US notice before any US launch.
Ownership warranty, no sharing, no presenting as real speech, we may refuse or delete, not medical care, UK law as default for the UK product.
Look at our Nairobi paper, what we will measure (acceptability, distress), ethics protocol, no efficacy claim.
Private 18+, rights-holder solo file, labelled stream, no download. Written AUP question. UK first, US later.
Stage 1: possibly free. Stage 2: possibly £6.99. Prices for later stages to follow. US pricing after US launch.
Research and collaboration: mental.health@ltns.cloud