How Amazon’s CodeWhisperer Aims to Rival GitHub Copilot in AI Coding

At the re:Mars conference in Las Vegas, Amazon announced the preview of CodeWhisperer, an AI‑driven coding assistant positioned as a response to Microsoft’s GitHub Copilot.

CodeWhisperer, built on machine‑learning models trained on billions of lines of open‑source, Amazon‑internal, API‑documentation, and forum code, provides syntax‑correct suggestions that adapt to a developer’s style, comments, cursor location, and surrounding code.

AWS chief evangelist Jeff Barr highlighted that the service aims to boost developer productivity and is free during the preview, while Copilot charges $10 per month.

The tool currently supports Python, Java, and JavaScript within VS Code, IntelliJ IDEA, PyCharm, WebStorm, and AWS Cloud9, and it can generate AWS‑specific snippets such as S3 bucket creation commands.

Amazon assures that, during the preview, developers’ code will not be used to further train the model, though it may be scanned after the preview period to improve performance.

Academic studies cited in the article note that similar AI assistants like Copilot have introduced security vulnerabilities and low‑quality code, suggesting that CodeWhisperer may face comparable challenges.

Amazon also offers CodeGuru, a cloud service that uses machine learning to detect vulnerabilities and automate code reviews, indicating a broader strategy to address code‑quality concerns.

Future plans include extending support to the AWS Lambda console, underscoring the growing competition between Amazon and Microsoft in AI‑assisted development tools.