TITAN (Translation and Interpreting Technologies and Accountability Network) investigates how artificial intelligence (AI)–based translation and interpreting (ATI) systems affect rights, democracy, and social well-being.
Our research combines translation and interpreting studies, law, public policy, sociology, technology, and gender studies to generate practical knowledge, guidelines, and a certification system for ethical ATI practices.
Why This Research Matters
AI-based translation and interpreting (ATI) systems are now used in hospitals, courtrooms, schools, government offices, and global companies. Their adoption is rapid, and the risks are expanding at the same pace. Among others:
- In healthcare, mistranslations of medical instructions can jeopardize patient safety.
- In law, biased outputs may reinforce gender stereotypes or cultural prejudice.
- In minority communities, poor language coverage accelerates the erosion of linguistic diversity.
- In politics, machine-generated misinformation undermines informed debate.
TITAN examines these risks to design safeguards that allow AI language technologies to be used responsibly, inclusively, and accountably.
Our Objectives
We structure our research around seven goals, each connected to specific actions:
- Classify risks and protections – compile a typology of ATI failures (e.g., medical mistranslations, biased asylum application terminology) and identify protective measures.
- Benchmark practices – survey organizations such as the WHO, ICAO, and the European Parliament to analyze strengths and weaknesses in their use of ATI.
- Advance ATI literacy – create training resources that help professionals and citizens detect unreliable outputs, such as errors in legal contracts or patient records.
- Monitor technological change – track new tools, including large language models, to anticipate their long-term social and professional implications.
- Foster international collaboration – connect researchers, policymakers, and industry (e.g., Disney, Lionbridge) to share evidence and set standards.
- Promote gender sensitivity and inclusion – document cases where systems misgender doctors, judges, or public officials, and design guidelines to reduce exclusion of women, minorities, and speakers of less widely used languages.
- Develop guidelines and certification – translate findings into practical standards and establish a certification scheme that signals accountable ATI use.
Our Approach
TITAN combines interdisciplinary research with active stakeholder participation:
- Systematic reviews to synthesize evidence on risks, rights violations, and protective strategies.
- Empirical studies including interviews, focus groups, and case studies in healthcare, law, education, and public administration.
- SWOT analyses to compare ATI adoption across organizations.
- Delphi studies where international experts refine classifications and recommendations.
- Consensus-building activities such as round tables with policymakers, NGOs, and industry partners.
- Training and literacy materials co-designed with users and made available in multiple languages.
Expected Impact
TITAN’s research delivers:
- For policymakers – guidelines to regulate ATI in ways that protect rights without blocking innovation.
- For industry – a certification scheme to demonstrate ethical and trustworthy practices.
- For professionals and citizens – training and resources to recognize risks and exercise their rights when using ATI.
- For society – protection of linguistic diversity, inclusive communication, and stronger democratic debate in an AI-driven world.
