So I asked ChatGPT a question about our profession...

[George Hatch]

Our board disciplined someone (in addition to other things) for not verifying the data with the county records.

I appraised a home where realist had the lot at 2 acres. Of course, the listing agent used 2 acres. Most appraisers use realist....

This county does not list the lot size for some homes, but does have a measuring tool. So I measured it. It came out to 1.5 acres, so I made the report subject to a survey. Survey said 1.5 acres.

My point is, it has to be 100% all or nothing in this new tech world. Same with regression and other tools.

If we use third party data, measurements, AI to wrangle the data, regression, etc. We should be free of liability from all sources....should not have to verify anything...eventhough the site size was incorrect. If I assume pdc cubicasa measurement is accurate, I can assume all other data sources are accurate.

If you have a known or readily knowable conflict then you have to make a choice between them. Either that or continue to pull on the dangling chad until you can make a more informed choice.

 

[DWiley]

And yet daily I see GSE employees opining that the accuracy of PDC based reports and software based condition and quality ratings are more accurate than those developed by appraisers.

Well, one CAN actually determine if a major crack, standing water, leaking roof, etc. that is visibly evident in photos was or was not reported.

 

[J Grant]

Well, one CAN actually determine if a major crack, standing water, leaking roof, etc. that is visibly evident in photos was or was not reported.

If an appraiser failed to report it on their own appraisal how do you know if that same appraiser would have reported it if a PDC was done instead on the same property?

 

[Essexfenwick]

If anyone really thinks mortgage appraisers will be around in 20 years is just silly.

5 years

 

[TerryRohrer]

Well, one CAN actually determine if a major crack, standing water, leaking roof, etc. that is visibly evident in photos was or was not reported.

For traditional appraisal reports, what is the error rate for identifying things that should be subject to repair or subject to inspection?

 

[CGinMN]

Identifying what constitutes a required repair is an opinion.

 

[Alsie35]

And they used it to drastically reduce the numbers of human generated appraisals.

This is a rare chance for appraisers to fight back and refuse to implement the new UAD. with no reliable appraisal the interest rates will go up and fewer buyers of automated bundles without a reliable system.

It's the only time I can think of where appraisers have power.

You go first.

 

[Alsie35]

If anyone really thinks mortgage appraisers will be around in 20 years is just silly.

Quite right, and not because anyone nefarious is plotting to "get rid of us".

"Its the technology, stupid".

AI + property inspection photos + scans will be generating value reports, based on a myriad of definitions and criteria, literally at the "push of a button" in the future. Appraisers' past collected data will be used to create those definitions and criteria, and after some prolonged testing period, they will - literally - no longer be needed. And we won't be the only profession that will "go the way of the Dodo" as a result of this new tech.

Now WHEN exactly all this takes place is probably not known for sure (just see how long it took the GSEs to do the UAD 3.6 changeover), but coming it is. But after probably after 90% of us on this forum today have retired.

That's one "positive", at least. But there's a lot of PITA with UAD 3.6 between "here" and "there".

As Lee Elia put it: "Its a disheartening ****ing situation we are in."

 

[George Hatch]

However long it takes on the CR side of the mortgage lending business the CG side of the business will probably only hold out for a couple more years beyond that. IMO.

 

[DWiley]

If an appraiser failed to report it on their own appraisal how do you know if that same appraiser would have reported it if a PDC was done instead on the same property?

I don't think anyone had made that assertion.