Amazon Looks to Attract Big Corporate Customers with New Chatbot and AI Protection Services

JAKARTA - Amazon is trying to attract large corporate customers to its AWS cloud computing service with a new chatbot for business. They also offer protection against legal and reputational harm that can arise from artificial intelligence results.

A new chatbot called Q is designed to aid productivity by helping workers summarize important documents, support tickets and chat via communications apps like Slack. Amazon said this at its annual cloud computing conference in Las Vegas on Tuesday, November 28. The software can also automatically make changes to a business's source code, speeding up development.

The new software comes about a year after OpenAI's ChatGPT hit the scene, sparking an investment frenzy in generative artificial intelligence startups. Alphabet and other companies have announced their own chatbots, which can talk like humans to help with everyday tasks.

AWS CEO Adam Selipsky, at Amazon's annual cloud computing conference in Las Vegas, announced new protections against objectionable content in generative artificial intelligence applications, called Guardrails for Bedrock. "This service allows users to filter out harmful content," he said.

Because generative artificial intelligence is trained with publicly available content, unacceptable words or other content may enter the results of user requests. This is especially problematic for young users, in situations of global conflict, or during elections where generative artificial intelligence results in search results can influence opinion.

Security advocates have warned that generative artificial intelligence could operate beyond human control and produce increasingly dangerous content or operate entire systems without oversight. Primarily, they worry that the software could place influential - and compelling - content on social media sites like X and Facebook.

Selipsky said the new service is important for customers to set boundaries appropriate to the generative artificial intelligence they use.

"For example, a bank could configure an online assistant not to provide investment advice," Selipsky said. "Or, to prevent inappropriate content, an e-commerce site could ensure that its online assistant does not use hate language or slurs."

As part of its appeal to companies, Amazon says its Q chatbot will offer businesses boundaries so it can keep sensitive data from employees who shouldn't have access to it. Prices will start from 20 US dollars (Rp. 300 thousand) per user per year.

Also at the conference, Amazon announced that it would provide coverage to its customers against lawsuits based on misuse of copyrighted material. For example, stock photography company Getty Images sued Stability AI earlier this year, accusing it of taking images from its website without permission.

"Guardrails for Bedrock is currently in limited preview," Amazon said. Unfortunately the Seattle-based company did not provide additional details about their warranty policy.