Appen launches three products for ethical and trustworthy generative AI

Appen Limited, the provider of high-quality data for the AI lifecycle, is unveiling three new products that will empower organizations to leverage the power of generative AI, as well as enhance the end-user experience with the expertise of diverse AI training specialists.

Appen’s launch of this series of products lends itself to the company’s expanding vision: an AI platform company that emphasizes client creation of AI-powered experiences for their customers, addressing complex data preparation, unusable data, incomplete data, and the lack of sophisticated data pipelines, according to the vendor.

Appen seeks to enable their clients to embark on a generative AI journey underpinned by their industry-leading research in transformer models and generative AI, according to the vendor.

"We are excited by this next phase for Appen. We will continue to create products and services that serve the data needs of our clients,” said Armughan Ahmad, CEO of Appen. “Our generative AI products we announced today are a great example, and we are just getting started.”

Ethics and AI are inevitably intertwined as the world moves forward to embracing AI as a powerful tool for a variety of use cases. The increased adoption of large language models has contributed to this relationship, raising questions of how to achieve ethically aligned and accurate human/machine interactions.

Appen’s Reinforcement Learning with Human Feedback, one of the recently launched products, addresses this necessary balance in AI; allowing users to create prompt-response pairs engineered by a diverse group of AI training specialists, Reinforcement Learning with Human Feedback reviews the pairs for accuracy and bias.

“Generative AI is going to change human experiences in ways that we have not even thought about. We at Appen are excited to power those experiences for our clients,” shared Sujatha Sagiraju, chief product officer at Appen. “It is critical to have human feedback while building generative AI applications to ensure brand integrity and prevent bias and toxicity.”

Appen’s Document Intelligence product, the next of the aforementioned releases from Appen, targets insight extraction from unstructured data existing as scanned and photographed documents. The solution allows unstructured data to become a usable data source, with accuracy results of 99% on diverse documents, according to the vendor. Ultimately, Appen’s Document Intelligence product will drive enterprise understanding of customer needs, preferences, and behaviors to better personalize interactions.

With Appen’s Automated NLP Labeling product, the last of the product releases, users can annotate data using zero-shot or few-shot learning techniques and large language models. The annotations are further reviews by AI training specialists to assure the dataset’s accuracy and quality.

“We strive to ‘Do Good, Be Good, and Lead Good,’” said Ahmad. “By ‘doing good,’ we aim to contribute to society through our crowd model and social impact initiatives. By ‘being good,’ we act with integrity and do the right thing as a company. And by ‘leading good,’ we strive to improve how AI impacts the world—ethically and without bias.”

To learn more about Appen’s ethical AI products, please visit https://appen.com/.