Re:Invent 2023: Chatbot Q and other AI-supported applications for AWS

It's no secret that Amazon wants to compete with the top dog of chatbots, ChatGPT. Now this competition also has a name: At Re:Invent 2023, AWS announced several products that are powered by generative AI. These include the chatbot Q.

What chatbot Q can do and why it has already been criticized

With the Q chatbot, AWS aims to simplify typical office tasks in particular. These include summarizing documents, answering questions or making suggestions for social media campaigns. As is usual for a chatbot, Q uses natural language for this.

One major advantage is the simplified use of the cloud itself. Q helps users to find their way around the cloud and provides information about the applications. This is certainly helpful given the mass of services and functions.

Chatbot Q can currently access around 40 company applications via interfaces, including applications from Microsoft, Google, Servicenow, Salesforce and, of course, Amazon.

To avoid misinformation as far as possible, Chatbot Q accesses the content of business customers within the cloud. AWS ensures that this customer data is never used for further training of the AI.

Nevertheless, there is already criticism of Chatbot Q. Like Platformer reports, some AWS employees have warned of risks. Chatbot Q is said to be struggling with hallucinations. The AI also passes on confidential information.

Amazon itself denies that there are security problems with Q. The news could hardly come at a worse time. After all, Amazon explicitly promotes its own chatbot as a more secure option compared to ChatGPT and other AI chatbots.

Amazon is already lagging behind in the race for generative AI and does not want to make a faux pas here.

Chatbot Q is currently available in a preview version, which starts with a free trial phase and can then be used with prices starting at 40 US dollars per month.

Transcription Platform and Titan Image Generator

Although the great hype surrounding generative AI was triggered by a chatbot (ChatGPT), other areas are also gaining ground. At the forefront is image generation using generative AI via applications such as DALL-E or Adobe Firefly.

Here, too, Amazon still has some catching up to do. The Titan Image Generator aims to close this gap. The AI uses natural language to generate images in many styles. Although the Titan Image Generator currently only understands English, other languages are sure to follow soon.

The Transcription Platform already contains over 100 languages and is now being updated with generative AI. This should make the transcriptions even better and more accurate. The system learns people's language behaviour using its own algorithms and can therefore better recognize individual patterns. This increases accuracy by up to 50 percent.

Operating AWS without chatbot Q with skill it

The Q chatbot does take some of the work out of using Amazon Web Services (AWS), but we can't rely on it completely. First we have to learn how to use and navigate AWS, only then will the support of a chatbot like Q be helpful. Finally, we have to tell the AI what we want it to do.

In our seminars, you will learn how to use AWS and develop your own solutions. In the course AWS Cloud Practitioner Essentials course will give you a basic understanding of AWS. In our three-day training Developing Serverless Solutions on AWS we use practical exercises to show you how to develop solutions within AWS. And all without the chatbot Q.