Time for some Q & A

The alpha release of laterite.ai is currently invitation-only as we are in the process of testing and fine-tuning our models and the functionality of the platform. In the next stage, we will be inviting researchers who have expressed interest in joining the platform to participate in testing. If you would like to be considered for an invitation to join the alpha release, please click the provided link to express your interest.

Laterite.ai applications are primarily intended for researchers in the social sciences, including both qualitative and quantitative researchers. This could include academics, students, researchers working in universities, think tanks, or other institutions, as well as people working in research or survey firms. The goal of these applications is to save time for researchers by automating tasks that can be time-consuming, such as preparing survey instruments, coding qualitative data or structuring the quant analysis. They can also be useful for brainstorming and idea generation. It is important to ensure that the input provided by users is in the proper format and instructions are clear in order to get the most out of these applications, but no specific skills are required beyond that.

Our goal is to provide useful applications to researchers that save them time and effort. Some of these applications are powered by artificial intelligence, such as large language models, which means they are probabilistic and can sometimes make mistakes or produce incomplete, incoherent or incorrect content. While this is a possibility, we hope that these applications can still be helpful to researchers by speeding up tedious tasks. We are constantly working to improve the quality and reliability of these models and welcome feedback from users to help us improve their experience.

When you upload text or data to laterite.ai, it is first sent to a server where it is processed using some automated scripts. If necessary, the server communicates with a large language model provider, such as OpenAI's GPT-3 platform, using an API to send snippets of text and receive a response. The data is then processed and combined into a final output that is sent back to you. Laterite will keep a log of interactions with the AI, including prompts and completions, for a period of 3 months in order to improve the quality of our models and ensure that the terms of use are being respected. The data will then be deleted. Data that is sent to the large language model provider will be treated in a similar manner, although logs may be kept for a longer period of time. To protect your privacy and the privacy of the people the data is about, we ask that you only upload data that does not include personal identifiers, such as names, telephone numbers, detailed location data, and ID numbers. Excluding any personal identifiers from the data you upload is a key condition in the terms of use. We are working on building automatic filters to remove personal identifiers, but these models are not yet ready.

Laterite.ai is continuously expanding the range of applications we offer to researchers. Our current batch of applications can facilitate the process of coding a survey using Open Data Kit (ODK), including using a natural language to ODK coder, automatically checking for errors in a surveyCTO form, translating a surveyCTO form into a word document output, or extracting questions from a survey instrument in word. Our models can also help with research instruments, including checking for biases and proposing better formulations, checking the quality of translations and back-translations, and highlighting risks. We also have applications that can assist with qualitative tasks, such as clustering and labeling a list of open-ended responses or classifying jobs by ISIC category based on a list of open-ended job descriptions. Additionally, we have applications that can help with brainstorming and problem-solving, such as by automatically creating an issues tree.

Our vision for laterite.ai is to become the go-to place for researchers looking for tools and ideas that save time, facilitate brainstorming and ultimately improve the quality of research. We are continuously blown away by the quality of tools and ideas that people come up with and are looking to work with researchers who have developed fantastic tools that we can integrate into our platform. In order to be considered as a contributor, your app should be written in Python, have a relatively light processing requirement, and produce a professional output. Ideally, these apps should build on large language models or other types of foundational AI models in order to enhance the quality of research. As an contributor, you will receive the majority of the net revenue generated by your application after all costs have been accounted for. The hope is that we can help you with all the back-end required to make this app available to a larger audience. If you are interested in becoming a contributor and contributing to our mission of improving research quality, please get in touch with us to discuss your idea.