Legal Question in Intellectual Property in California

Sounds more like a derivative work to me....the reviews being excerpted and edited are not mere data, they are individual creative works.

To determine whether the resulting work of company A's algorithm is a derivative work from the original and thereby constituting an infringement requires a factual analysis of the resulting output. Copyright infringement and its defenses is generally a balancing test of many factors, without a single element being determinative.

If the output to your algorithm is merely a statistical result (e.g. result showing that 80% of internet users like product X), you will likely be okay. However, if the result is composed of portions of the user's review, you might be in dangerous waters.

On top of the possible copyright issue, you should also consider violation of the Terms of Use agreement with the data source website and potential issues of privacy violation if your algorithm also retrieves user identifiable information.

My advise would be to consult with an attorney to review your situation in greater detail.

Kind regards,

Jim