Odd. The market leaders in AV technology all use cameras to read road signs... The problem with a database is it's useless unless it is updated in real-time, with no errors. With static signs, this is maybe a believable possibility, but with a significant chance of bad data caused by human error. With temporary signs, there is no chance of good data. Too many individuals and organisations use them, sometimes for very short periods of time. There is no way every one could be relied upon to supply data that is accurate and up-to-date, every time. If AVs need to use a camera to read temporary signs, the system is still open to abuse. If both a database and a camera are used, what happens when they disagree? Do you assume the database is accurate, ignoring the possibility of bad, out-dated, or missing data? Or do you assume the road sign is correct? This isn't something I made up for a forum post. It's a well-known and we'll documented challenge in the development of AVs. They can (and do) misread road signs due to vandalism, damage, overgrown foliage, and a whole host of other hazards. There is every chance that this could be exploited for malicious purposes.