In part two of a two-part series, the Balchem technical team selected industry research of interest from the 2024 American Dairy Science Association meetings to feature on this episode of the Real Science Exchange.
In part two of a two-part series, the Balchem technical team selected industry research of interest from the 2024 American Dairy Science Association meetings to feature on this episode of the Real Science Exchange.
Smart Cows, Smart Farms: Unleashing the Potential of Artificial Intelligence in the Dairy Sector
Guest: Dr. Jeffrey Bewley, Holstein Association USA (1:58)
Dr. Bewley is the Dairy Analytics and Innovation Scientist at Holstein Association USA, where part of his role is collaborating with Western Kentucky University at the WKU Smart Holstein Lab. The group works with more than 30 technologies, including wearable, camera and machine vision, milk analysis, and automation technologies. At ADSA, Dr. Bewley’s presentation was part of a symposium titled “Applications of AI to Dairy Systems.” His talk focused on cow- and farm-level technologies using artificial intelligence. He anticipates a continued massive increase in the availability of technologies for dairy farms to assist with automating processes that are often monotonous tasks. One example of this is the wearable accelerometer technologies that allow for the assessment of estrous behavior, as well as rumination and eating behavior. In the future, camera-based technologies may become more commonplace for things like body condition scoring. Cameras may also be able to monitor rumination and eating behavior, and even perhaps dry matter intake. Dr. Bewley also sees an opportunity on the milk analysis side to be able to measure even more biomarkers to better manage for improved health, reproduction, and well-being. He reminds listeners that animal husbandry will continue to be a critical piece of dairy farming even with advancing technology. He gives examples of current and cutting-edge technologies on the horizon for dairy farms. On his wish list of technologies for the future, he includes dry matter intake measurement and inline measurement of somatic cell count, hormones, and metabolites in the milk. In closing, Dr. Bewley encourages listeners to be excited yet cautious about artificial intelligence and gives examples of how technology can collect phenotypic data to use in genetic evaluation.
Explaining the Five Domains and Using Behavioral Measures in Commercial Systems
Guest: Dr. Temple Grandin, Colorado State University (26:48)
Dr. Grandin’s presentation was also part of a symposium, titled “The Animal Behavior and Wealthbeing Symposia: Evaluating Animal Comfort and Wellbeing Using the Five Domains.” The five domains approach is gaining popularity. Previous guidance documents emphasized preventing suffering, cruelty, and discomfort. The five domains are nutrition, environment, health, behavior interactions, and the emotional state of the animal. Much of the information available is very theoretical. Dr. Grandin’s goal for this presentation was to gather easy-to-download scoring tools to assist in auditing the five domains in the field. She emphasizes the importance of good stockmanship for animal well-being and cautions that while artificial intelligence technologies can be used to assess the five domains, good stockmanship will always be necessary. Dr. Grandin recommends a three-legged audit: internal, independent third-party, and corporate representatives. She cautions against farming all audits out to a third party and anticipates that it has the potential to cause major supply chain disruptions. Lastly, Dr. Grandin recommends simple yet effective outcome measures for audits that can be taught in a short training session that includes practice audits.
View her five domains paper here: https://pubmed.ncbi.nlm.nih.gov/36290216/
Please subscribe and share with your industry friends to invite more people to join us at the Real Science Exchange virtual pub table.
If you want one of our Real Science Exchange t-shirts, screenshot your rating, review, or subscription, and email a picture to anh.marketing@balchem.com. Include your size and mailing address, and we’ll mail you a shirt.
Scott Sorrell (00:07):
Good evening everyone, and welcome to the Real Science Exchange, the pubcast where leading scientists and industry professionals meet over a few drinks to discuss the latest ideas and trends in animal nutrition. Now, tonight's gonna be a little bit different. In this episode, we're gonna be reviewing some presentations that were given at this year's ADSA. Now, the problem with the ADSA is there's such great content and, and presentations there that you just can't get to all of 'em. So with that in mind, we ask our technical team to scour the booklet and come up with some of the best content out there that we could share with you, our audience. And so that's what we're gonna be doing in this episode. We're gonna be going over different presentations and papers that were presented at this year's ADSA for your Enjoyment. So enjoy
Commercial (01:02):
The economics of feeding ReaShure Precision Release Choline. ReaShure is fed during the transition period, and because it's fed for such a short period of time, it costs just $15 per cow. And yet the benefits will continue to generate income throughout the year. Cows fed reassure produce five pounds more colostrum, which pays for your reassure investment on the very first day of lactation cows. Fed Reassure also produce five pounds more milk per day every day. That means after the first day, every day is payday invest and reassure during the transition period and recoup your investment on the very first day of lactation. After that, you got it payday. And we're here with Dr. Jeff Bewley from the American Holstein Association.
Scott Sorrell
Now, Jeff and I, we've known each other for quite a while, and so I'm looking forward today's conversation. Jeff, before we kind of get started into your presentation, would you mind telling us a little bit about yourself?
Dr. Jeff Bewley (02:16):
Sure. I grew up on a dairy farm in Kentucky, and pursued education in that area at the University of Kentucky. My master's from the University of Wisconsin and PhD from Purdue. I've worked at a few different places over the years. I've worked in extension at the University of Kentucky. I've worked for Bova Sink and Alltech and ICE Robotics and Performance Nutrition, and I've been with Holstein Association USA now for the last four years as the dairy analytics and innovation scientist.
Scott Sorrell (02:49):
Okay. And then you also have a relationship with Western Kentucky University, something called the Smart Holstein Lab. What, what in the world's a smart Holstein lab?
Dr. Jeff Bewley (03:01):
Sure. So a few years ago we came up with the idea of working with Western Kentucky University and, and their dairy, and a collaboration that we call the WKU Smart Holstein Lab. And this is a center for exploration of technologies in the dairy industry. We work with a lot of different technologies. I think last count over 30 different technologies that we have there at the Smart Holstein Lab. We work with wearable technologies. We work with camera and machine vision based technologies, milk analysis technologies, and automation technologies there.
Scott Sorrell (03:42):
Oh, excellent. So that's a great segue into the presentation that you gave. Now, your presentation was part of a symposia there at the ADSA and that was titled Applications of AI to Dairy Systems. And then your talk specifically was smart cows, smart Farms, unleashing the potential of artificial intelligence in the dairy sector. So why, just to kind of get us started, give us kind of an overview of what your presentation was all about.
Dr. Jeff Bewley (04:13):
Sure. I really enjoyed the session and hearing from the other speakers also, but my focus was primarily focused on cow or farm level technologies using artificial intelligence. And so I talked a lot about the types of technologies that we have available today and where we're headed as an industry in, in terms of using artificial intelligence on farms is really exciting when you think about the first things that we've started seeing on farms now and more exciting as we think about what's coming in the future with artificial intelligence. And I think so much of it goes back to what we start to see in our world today. If you're using chat GPT in any format so far, it's really amazing what chat GPT by itself can do. And it gives you ideas to start thinking about what we can do with artificial intelligence on farm also.
Scott Sorrell (05:14):
So you, you talked about what's coming and so maybe it's a little early to jump into that, but I, but I'm, I'm always curious, what, what's the dairy farm gonna look like in 2050? How different's gonna be from what we've got today?
Dr. Jeff Bewley (05:29):
To some degree, I will say it's hard to say for sure because so many of the neat technologies that we have on our dairy farms are where we basically borrow ideas from other industries. So some technology is developed for the human wearable industry or for the automobile industry or, or automation in another area. And then we're able to take that base technology and, and apply it to the dairy industry. So it, it's hard to project exactly where we'll be, but I think that what we're going to see is, is a continued massive increase in the availability of technologies for dairy farms to help with automating processes, automating tasks that sometimes are, are monotonous tasks that we can automate to help manage the farm or manage the cows, and then to monitor the animals. Also, the, the best example of that at this point is the wearable technologies where we basically took an idea from a technology called an accelerometer and accelerometer measure motion in three dimensions.
Dr. Jeff Bewley (06:40):
It's the same technology that's in our smartphone, that's in Fitbits, it's in a lot of automobiles. And we take that idea and we put it into a device that we attach to or put into the animal, and we measure motion so that we can pick up Esther's behavior or rumination behavior, eating behavior, et cetera. So that's, that's the best example of, of A to Z where that technology came from an idea to where now it's really, that's a pretty mature set of technologies that work very well, and it's advanced because we were able to borrow that idea from those other industries where we're headed. I think we'll see a lot more technologies that use camera based technologies. So, you know, if I go to the airport now, they do facial recognition to identify me and that based technology we're taking and applying it to dairy animals so that we can do things like body condition scoring or, or locomotion scoring of animals.
Dr. Jeff Bewley (07:41):
I think we'll also get to the point where we're able to do more identification with that type of technology and, and in the future, even where the cameras themselves are monitoring things like behaviors of animals. So they're able, they will be able to monitor rumination behavior and eating behavior and even perhaps dry matter intake from a camera-based system as opposed to having a device that's attached to every animal. And then the other area that I see a huge opportunity for on the monitoring side is in milk analysis. So I look at the, the milking parlor and biologically we're collecting samples from our animals three times a day or, or more in robot situations. And we're able to collect those samples and measure some things already. Things we've thought about before, like fat protein and somatic cell count. But moving forward, we're going to find more biomarkers that we're going to be able to identify to better understand what's going on with our animal physiologically so that we can better manage that animal for improved health, improved reproduction, improved animal wellbeing.
Scott Sorrell (08:54):
So you've talked a lot about new ways to monitor the animal, right? And, and, and I've got a bias that today we've already got a lot of monitoring and a lot of data, and maybe, maybe we're not doing a great job of synthesizing that data and making decisions with that. Can you talk a little bit about how, how AI then is gonna take the, the data we currently have and all the new data from all the new monitoring that we're gonna have, and, and what does decision making look like then with all this new information?
Dr. Jeff Bewley (09:25):
I think that's a great point. I, we have a lot of information already. Even without buying many technologies, we have a, a lot of information. And you can imagine with all the technologies that I have, how many how much information overload that I, I go through AI will help simplify that. So one of the things artificial intelligence can do is look for patterns in data that process data more like what our human brain does, where we're pulling in lots of information from lots of sources to make decisions or form opinions. And AI helps to, to mimic that process as opposed to traditional statistics, which is what most of us were trained in, is traditional statistics where you maybe only look at at one thing at a time. And we can pick up these patterns with artificial intelligence, can do it faster many times than what the, the old way could do and can identify things that we wouldn't necessarily see otherwise.
Dr. Jeff Bewley (10:26):
That all comes with a set of cautions, of course. We have to recognize that, that it is a machine and sometimes it can be wrong. And we have to, I think at the base we need to understand the biology of the animals that we're working with and the system that we're working with. So there's still always a need to understand the cows and understand the biology around the system that we, we work with. It's not just that we, we sit and we look at, at a computer all day. There has to be some basic understanding going back to that.
Scott Sorrell (10:59):
I'm reminded of a far side comic. And this goes back years. And, and it was, if I remember correctly, it was like a big computer and there was a, a person in the picture as well as a dog. And the question was, okay, so the computer's running the operation and the man standing there sitting and they're like, so what's the dog for? Well, it's to keep the man from touching the computer or something like that. So I'm just kinda, I know it didn't tell real well, but I'm kinda wondering, right, what is the role of the dairy farmer gonna be you know, when once AI takes hold? I don't know if you have any thoughts relative to that, but Yeah.
Dr. Jeff Bewley (11:42):
Well, I think to some degree it's focusing more on, on some of the management tasks. I think there's still that interaction with the animal that's important. And I think that it's, it's not going to overtake everything. It's, it's only going to help us. It's a tool, not a replacement for all the things that, that we've done. I think it, it will help make our jobs easier in some ways and perhaps harder in other ways, but it's not a replacement for people in total. It is in some parts though, that, that AI is able to do things that perhaps we couldn't do. Just yesterday, out of curiosity, I uploaded a ADH, I report into chat GPT and asked it to give me a report of the, of the trends and what it saw as opportunities. And, and it took it 15, 20 seconds. And I was really amazed at, at what it was able to do to summarize that DHI report, well, that's a job that, that people in the field do all the time. So maybe some of those things can be done automatically as opposed to spending hours and hours looking through reports.
Dr. Jeff Elliott (12:57):
So Jeff, we're, we're seeing, just like we're doing here, we're seeing a lot of presentations and seminars given on, you know, precision technology. Do you have any idea what the actual uptake is on dairies? I mean, are a lot of them trying something or a lot of them scared to?
Dr. Jeff Bewley (13:24):
I think it depends on the definition of precision. But if it took just the wearables as an example, I, I saw a survey the other day that that indicated that, that it may be around half of dairies that are adopting those technologies at this point. For, to me, this is, the wearables are, are old now because they've been around for 15 years. So it, it's not that that happened overnight. It, these technologies had to work through some challenges and so forth to get to the point where they are now. But it's also not where 100% of producers have that particular technology. I think for the most part, most dairy producers would have some kind of a technology. It may not be that one, but would have some kind of technology, the robots, it's still really pretty low in, in the US in terms of adoption rate.
Dr. Jeff Bewley (14:16):
And I think one of the most important things that, that I think about a lot with regard to technology adoption, I get the question a lot, well, what can we do to get more people to invest in technologies? Or what can we do to make sure that everybody has a technology or everybody has this particular type of technology? And to me, that's the wrong question. The question should be, is there a need for something on the farm that can be fulfilled with a technology? So for example, if I'm having an issue with, with heat detection, perhaps I should look at and technology for heat detection, where if I'm having an issue with lameness, perhaps I should look at a lameness detection technology. But also it, it's important for the dairy producer to understand, you know, they're, they're looking at an investment portfolio of things that they can invest in. And sometimes in that scenario, for example, we're, we're looking at lame detection, maybe it's not that they should invest in a lameness detection technology. Maybe they should invest in improving cow comfort to prevent the cows from becoming lame in the first place. So to me, the question always comes to what is the need of the individual farm? Not how can we get adoption rates to 80% or 90% or whatever.
Dr. Jeff Elliott (15:38):
Okay. So I think that may go into another question I had in your abstract. You'd mentioned often there's over marketing of how AI is being employed. So is that kind of what you're referring to?
Dr. Jeff Bewley (15:54):
I think that's a little bit different issue, but what I see happen a lot of times is people say we're using artificial intelligence, and in actuality they're, they're really not using artificial intelligence. They're, they're just basically looking at numbers, standard deviations from a mean or something like that. That's not really artificial intelligence in the, in the truest fashion. And the other thing is that when we're using artificial intelligence, we have to recognize that in order for the system, you hear people say, well, it keeps improving over time because it's, it's learning. Well, it only learns if we provided feedback. So for example, if, if I have a system that's supposed to be detecting cows with metabolic disorders, and I never tell it which cows actually have metabolic disorders, it doesn't have anything to learn from the, the, the way it learns is that say, this cow did have milk fever, this cow didn't, or this cow did have ketosis and this one didn't.
Dr. Jeff Bewley (16:52):
So that continuous feedback loop is really critical for continual artificial intelligence improvements. The other thing is that I hear people say, well, we used artificial intelligence and this is, so it's better, it's not necessarily better. It can be better, but ultimately we still need to hold artificial intelligence up to the same standards that we would any other type of a calculation. If I have something that predicts something 50% of the time, it's still just a coin flip. Even if it use artificial intelligence, it's not particularly useful if it's only right half of the time.
Dr. Jeff Elliott (17:34):
Yeah. Interesting. And, and this may be a repeat, but my computer glitched a little bit earlier. Did you refer to maybe using the cameras on measuring dry matter intake?
Dr. Jeff Bewley (17:52):
Yes.
Dr. Jeff Elliott (17:53):
To me that's very important. 'cause You went on to mention about how we can monitor milk. We get all the milk to, you know, test a milk and daily milk weights, and if we can get that dry matter intake, we can really hone in on feed deficiency from a daily standpoint. So I just wanna confirm where we're at with me getting those dry matter intake measurements.
Dr. Jeff Bewley (18:15):
There, there is a technology on the market now that does dry matter intake using cameras. I think it's fairly accurate. The challenge with it is it requires a lot of cameras to be able to do it. So there's an economic cost to that too. And I may, maybe that's another point to bring up here is that we can be technically as, as interesting or useful as we wanna be, but ultimately at the farm level, it has to make economic sense to be able to invest in the technology. It has to be able to provide a return. And there's some examples now of technologies that were very technically correct and useful and interesting, but the reason that they were not adopted is they were just too costly to justify investment in that technology.
Dr. Jeff Elliott (19:01):
So on the, on some of these farms where maybe they'd been reluctant to try something and they finally did, and I think, I think you mentioned heat detection or, you know, herds that were using tail paint and then, then they decided to go into the activity monitors. Are those herds for a while maintaining both procedures so that they can check, you know, they're probably comfortable with, you know, tail paint, but they're wanting to go to these activity model. How do they, what's that process look like?
Dr. Jeff Bewley (19:31):
Typically what I've seen is, is in that scenario, people don't believe it immediately, but it doesn't take very long on estrogen detection to figure out that really works. So when you start seeing that it's agreeing with what you see visually, whether that be tail paint or mounting cows, but also what the breeder feels whenever they, they see the tone to the uterus and everything when they, when they breed that cow, then you start to really believe, okay, this technology, it's real, it really does pick up on, on asterisk and asterisk works. It really works well. The, the increase in activity around heat is, is huge for, for cows. And that technology's very mature and there are many, many options on the market that work very, very well for Esther detection. Yeah.
Scott Sorrell (20:20):
Jeff, one of the things I was kind of curious about, you mentioned there at WKU that you have a lot of different technologies that you, that you have at your disposal and obviously measuring a lot of a lot of different things. Is there anything that you want to measure that you can't measure today? And what would that be?
Dr. Jeff Bewley (20:39):
Good question. I still can't measure dry matter intake. To me, that one is one that I've always wanted to, to measure, measure. That's, that's probably the number one to me is dry matter intake. I'd also like to be able to get to the point where a lot of these milk-based technologies become more mature. So I would really like to be able to get inline somatic cell count working well, there's multiple companies working on that now, but where I can have somatic cell count for every animal, every milking, I'd like to see more perhaps in terms of, of reproductive hormone analysis in the milk. So if I can do estro detection and reproductive cycle management through the milk as opposed to through a wearable technology, or to be able to go more specific about the physiology of the animal as opposed to an indirect indicator.
Dr. Jeff Bewley (21:38):
So rumination behavior, that's, that's a very useful technology. But when we have a cow that drops in rumination, we generally know that something's wrong with the animal. We just don't know what's wrong with the animal. But if we're able to measure BHBA levels, for example, or calcium levels in the milk, or perhaps with a wearable technology that works more like the glucose monitors that people wear sometimes, then if we're able to do that, then we have a, a diagnosis of what's going on with that animal, which takes it to a whole additional level beyond just knowing that something's wrong with the animal. Now I can say this animal has mastitis, this animal has ketosis, this animal has milk fever. That to me is, is the ultimate where I'd like to see this go. And I, and I think in time it will get there.
Scott Sorrell (22:31):
Yeah. And how much time, what's, what's that look like?
Dr. Jeff Bewley (22:36):
I don't know.
Scott Sorrell (22:37):
Yeah, things are going awful fast. You had mentioned before that you'd like to see more ability to measure hormones in the milk, those kinds of things. What's keeping from that from us today? Is it simply adoption? Is it cost? Is it the technology's just not there yet?
Dr. Jeff Bewley (22:56):
Some of all the above. One of the issues, if we really wanna measure hormones in the traditional way, there's reagents involved. So there's a variable cost associated with using that technology. And so cost becomes the issue. But some of the technologies that are using infrared and light scattering approaches may be able to, to examine those things indirectly, again, using artificial intelligence so that there's no reagents or no variable cost associated with using that system. And I, I think that's probably where a lot of that's headed.
Scott Sorrell (23:32):
Yeah. Very well. Jeff, anything else from you?
Dr. Jeff Elliott (23:37):
Not from me. This, it's, it's highly interesting and it's just, it's gonna be interesting to see where we go in the future with this,
Scott Sorrell (23:44):
For sure. Yeah. Yeah. I agree, Jeff. Very interesting. And, and Dr. Ley, I just kind of ask in closing, any, any final comments or thoughts key takeaways that you'd like to share with the audience?
Dr. Jeff Bewley (23:56):
I think we should be excited about artificial intelligence. We have to be cautious about it. We have to ask good questions when we're thinking about investing in technologies or, or using new techniques on farms. But we need to be excited. There's a lot of opportunities. There is that point of information overload. We all have to recognize that you can't really run a dairy the way that we are running the technology, all the technologies we have. It doesn't make economic sense to do that. The other piece that I, I didn't mention yet that I see so much potential for is that a lot of these technologies provide us with new data that we haven't had access to before. That gives us a phenotype that we will be able to use in genetic evaluations moving forward so that we can select for animals that do things that we'd like for them to do.
Dr. Jeff Bewley (24:48):
So, for example, I think about body condition score, and we have a few body condition score technologies out there now. Well, there's a genetic component to the shape of the body condition score curve. Some animals are naturally predisposed to losing more body condition score than others, and that's a heritability of around 20%. Well, in the past we've not been able to select for that because we didn't have the phenotype. But now with this type of technology, then we have a phenotype that can select for an animal that better maintains their body condition score, which ends up being a, a healthier, more profitable animal. Then we can include that in genetic evaluations to select for animals that do what we'd like for them to do. So that's really, and then, then Jeff mentioned the feed deficiency. Yeah. You think about feed deficiency as a nutritionist, I think about nutritionist. I think about it from a genetics perspective. If we have drought matter intake data on more animals, then that improves the data that we have to be able to select for animals that are more feed deficient. So to me, although it's not the direct application for many of these technologies, the indirect application of using this data for genetic evaluations is, is also equally exciting. Mm.
Scott Sorrell (26:06):
Yeah. Excellent point. Well, listen, it’s all very interesting. Fascinating, in fact, and I want to thank you for joining us today. Jeff, this has been great. Thank you very much.
Dr. Jeff Bewley (26:16):
Thank you all for having me. All right.
Dr. Jeff Elliott (26:18):
Thanks.
Commercial (26:20):
AminoShure XM Precision Release Methionine, the next generation in amino acid balancing with AminoShure XM. You can save up to 5 cents per cal per day on your methionine investment. Try it today and receive an additional 2.50 cents per cal per day savings with Balchem’s Limited time rebate offer. Contact your Balchem representative to learn more.
Scott Sorrell (26:48):
Right. And we're back with someone that really needs no introduction. And that's Dr. Temple Grandin. Dr. Grandin gave a presentation at this past ADSA 2024 during a symposia called The Animal Behavior and Wealthbeing Symposia Evaluating Animal Comfort and Wellbeing using the five domains. And then within that symposia Dr. Grandin understand that you gave a presentation then that was titled Explaining the Five Domains and using behavioral measures in commercial systems. And so I'm looking forward to having a nice conversation around that. But so why don't we jump right into it. Give us kind of an overview. You know, I've never heard of the five domains before, so maybe that's a great place to start. Can you talk us through kind of the, the overview of those? Well,
Dr. Temple Grandin (27:39):
The five domains is getting the five domains is getting more popular because previous guidance documents emphasized preventing suffering, preventing discomfort. And it's super important to prevent cruelty and discomfort. But what about having some, the fifth domain being, you know, the emotional state is the animal having positive things Like if you see a dairy cow using one of those brushes, that's totally positive. Now, I've done a lot of work on training animal welfare auditors out in the field, and you have to have very clear, simple guidance. 'cause We have to train 'em in a two to three day workshop. And then they have two practice audits with an experienced auditor you can't, you're not gonna have a super trained behavior person. So less guidance is very clear. It's hard to implement out in the field. So what I've tried to do in this paper, I also wrote too, present it in a simple way.
Dr. Temple Grandin (28:38):
Also, I referenced a lot of easy to download scoring tools. Okay. For example, in the five domains is nutrition, environment, health, behavioral interactions. And then the fifth domain well, in nutrition, one of the things you'd measure is body condition score. So where can people find good body condition scoring charts? Emphasize the ones that are free online as a few behind a paywall had to use, because the way I look at it is, it's an international kind of guidance thing. It's a paper I did for animals. I know that's not your journal, but it's very open access. And I'm very, very interested in getting things you know, spread around the world.
Scott Sorrell (29:28):
So real quick, again, the five were nutrition. What was the second?
Dr. Temple Grandin (29:33):
Environment, health, behavioral interactions. And then finally the fifth domain. Now, one thing that's very similar with the five domains is it's very similar to animal welfare quality from Europe nutrition. They call it good feeding environment. They call it good housing. Then you have the health and you'd have behavior. Now the thing is different is what about the fifth domain, which is the animal's emotional state. Now the way the five domains is presented, they say, well, you can't directly measure that because then I have to put something painful like castration into health, which to me doesn't make very much sense. We know that hurts. Yeah. I originally put that in the fifth domain as, as something bad. And I put something like the dairy, the motorized brush that dairy, I know I'm not supposed to say they love it, but dairy cows love this thing as something positive.
Dr. Temple Grandin (30:25):
And we play games with our horses the big ball down in Denver, Uhhuh, they love it. Cool. you see, those would be things that are positive. But a lot of the information that's available on this is very theoretical. And I'm gonna just present something. This is stuff that people can use out in the field. Right. And I'm getting ready to write some other stuff. And I think I'm gonna say that I think it's kind of odd to put things that hurt like dehorning and castration into health, but that's where they, where Mal David Malaria wrote the fifth five domains puts it. But the one reason why I think a lot of commercial companies are going for this is they like the idea of, of a continuous trait. How could you have on the old, old freedom, how can you have total freedom from discomfort? I've been on a long airplane trip, yeah. How, how about a trip to Australia and smoking in the back of a 747?
Scott Sorrell (31:27):
A lot of discomfort there.
Dr. Temple Grandin (31:28):
Rock. Oh, I had big discomfort on that flight. It was a horrible flight. I was suffocated from cigarette smoke. That was back in the seventies. I didn't have freedom from discomfort on that flight. Ugh. Like 12 hours or something. Terrible.
Scott Sorrell (31:43):
You're saying that freedom from discomfort's kind of an unattainable goal.
Dr. Temple Grandin (31:48):
Well, you can't have total freedom from anything. So like a lot of people like the idea of a, of a continuous trait. Yes, I can do body condition scoring, I can score lameness in cattle, in other animals. That's, you know, from mild lameness to very severe lameness. Even that is, you know, on, on a continuum. I can score you know fighting behavior between animals that would be something I don't really want. But all of these things kind of happen on a continuum, and that's something that a lot of the commercial people liked that factor. What I did in this paper is for dairy cows and beef cattle, pigs and chickens. I went and I looked up all the good scorecards I could find that people can use, like lay Well, for example, has a wonderful feather condition scoring for laying hands, and it's free. And they're beautiful scorecards. Animal welfare quality has some really beautiful scorecards for things like hernias and pigs, for example. But they have a way they calculate the final score. That's totally ridiculous. Also, they, they it, the, their audit takes too long to do, but they got some beautiful scorecards that I recommend people use for, for many of the different animals, especially on cleanliness of waterers.
Dr. Jeff Elliott (33:10):
So Temple, you've mentioned horses, you've mentioned poultry, you've mentioned pigs. So these five domains
Dr. Temple Grandin (33:18):
Would apply to any animal originally any animal. And originally it was developed for zoos. So yeah, any animal.
Dr. Jeff Elliott (33:26):
Oh, so they were using 'em in zoos. That's interesting. Yeah. Especially from a consumer or a customer that the general lay public's gonna be at the zoo. Yeah,
Dr. Temple Grandin (33:36):
Well, you know, when they, they that's where some of the research on that was first started, and then they obviously copied some stuff from animal welfare quality. Okay. Because the first four domains are pretty much the same, except in the behavior on the five domains. They put it into three categories in the behavior, behavior with other animals are they getting along with each other, licking each other or fighting how people, the stockmanship, I've done a lot of work on animal handling scoring systems where you score vocalization, pro use slips and falls when you're handling animals and then interaction with the environment, things in the pen, you know, they, the horse chewing it up or so there's three levels of behavior there. And, and good stockmanship really matters. And that would be the behavioral interaction with people, and that can be evaluated. And oftentimes the good stockmanship does not get enough credit. Right. People want the magic thing now. It's gonna be AI magical thing. It's gonna fix everything. Well, you have to also have the good stockmanship and it does matter. And there's a bunch of studies that show that animals that are afraid of people are less productive. And Paul Hemsworth did some of the very first work on that,
Dr. Jeff Elliott (34:56):
But, but one of our other ADSA reviews that we've done here has been on AI technology on dairy farms. Even though it's not gonna, like you said, replace, you still gotta have good stockmanship. Are there ways to use that AI to help further move these five domains and
Dr. Temple Grandin (35:20):
Well, I went to farm, farm in upstate New York that had an AI system that Merck was working on. And it even worked on the farmer's, ancient old computer and very simple to use. And it would pick out cows that had things wrong with him. And then he had to go and check 'em for mastitis, check their temperature. He also could find cows that needed breeding. And he loved it because he was able to, you know cut down on one person on heat detection and that stuff, and it was working really well for him as a subscription basis. I think there's some things that can be really helpful, but it tells you which cows to check. You still have to get the cows and check 'em. You still have to do that. It just tells you which ones. So instead of hunting through 200 cows in a pan, they'll tell you which ones to get. And he, and he was very, very happy with it. No, there's gonna be some good things going on with ai. You know, for, another thing that's being used for right now is things in chickens' lesions on feet. Instead of having a person score them a match a computer can very easily score a damage on, on chicken feet. Wow.
Scott Sorrell (36:36):
Dr. Grandin, you had obviously listened to all the presentations there at the ADSA. What would you say are some of the key takeaways from all the presentations? How would you summarize the information presented there?
Dr. Temple Grandin (36:49):
One of the things I, I remember in that meeting, and I think it was on a poster, is we've pushed the modern dairy cow to give so much milk that she's sort of like milking herself into the bucket. And that probably has something to do with why they don't breed very well. You see this, this gets into over selecting for a trait, where do you stop? Yes, we need high productive cows, but there's a point where you've pushed too far and probably need to stop. We're having problems right now in beef cattle with congestive heart failure now at lower altitudes and a lot of lameness things, a lot of crossed claws, which is genetic and it's related to selecting just for meat traits. This is something that 10 years ago just sort of had hints of starting. And certain places where, so like 30% of the Fed steers in APH have got some degree of heart swelling and congestive heart failure.
Dr. Temple Grandin (37:39):
That's just terrible. And that's published literature. It's totally published literature. You, you push a system too hard, you're going to start to get in. Yeah. Trouble. This is something I'm very concerned about. We're pushing for weaning, high weaning weight at a young age. And then some of these cattle have a, a twisted claw, which is a defect. Yeah, it's genetic. I've got a picture of a four month old heifer that hasn't eaten any grain. There's no way it's founder, it hasn't had any grain and it's got a twisted claw. You know, we're breeding cats that have got deformed legs as we think they're cute. I have problems with that. I'm gonna bash the pets just as much as, one of my big concerns right now in animal welfare is pushing the biological system to the point where, where the animal's got big problems, whether it's a pet, a bulldog that can't breathe or it's a farm animal.
Scott Sorrell (38:43):
Yeah. All things we humans have done to 'em through genetics. Dr. Grand, thinking about your presentation, what would you say are some of the key takeaways? One, some of the things that you wanted to leave with the audience?
Dr. Temple Grandin (38:55):
I wanna wanna lead with the audience that we can simplify this. So it's gonna be easy to train auditors in these short workshops that we have. That they need to be, you know, finding good visual scoring tools for the different things and, and get away from the theoretical stuff and how do we actually do it in the field. Okay.
Dr. Jeff Elliott (39:17):
And do you find that some of the, I guess the corporate buyers are kind of requiring more of these audits?
Dr. Temple Grandin (39:26):
Well, corporate, I do work with a lot of corporate buyers. They're my main clients right now. And I still do training sessions with corporate clients. I'm a Paco animal welfare trainer. I've trained hundreds of auditors now, and I've learned you have to make it very simple. So we're gonna get inter observer reliability. And so I really like clear outcome measures. We can measure lameness, we can measure skinny cows, we can measure swollen hots.
Dr. Jeff Elliott (39:54):
Are those auditors, are those auditors are they an employee of the, of that buyer let's say, or are they third party? Or is it a combination?
Dr. Temple Grandin (40:05):
They can be either. So you have people out there that work for third party auditing companies that are not employees, and then you have employees of the company that also are learning how to do this. Also, you teach the dairy too that do it. So I always say there's three legs on an audit, tripod, internal audits, third party independent, and what I call corporate or second party where the corporate people get out and inspect some stuff. You don't wanna just farm out out everything to third party. You are gonna end up having a supply chain wreck.
Scott Sorrell (40:39):
Anything else?
Dr. Temple Grandin (40:40):
No. I, I just you know, wanted to kind of do a presentation from the standpoint of, yes, I am an academic, I am a professor, but I'm also an auditor trainer. And do it from the standpoint of how am I gonna train somebody in two and a half day workshop and two practice audits to go out and do this stuff, otherwise it's just not gonna get done accurately. How do we simplify it but still make it really good, make it really effective? And I designed very simple scoring system with five key indicators for the meat industry, almost like 25 years ago now. It's been very effective, very simple, very effective outcome measures, insensibility stunning localization, pro use, falling
Dr. Jeff Elliott (41:23):
Simple.
Scott Sorrell (41:25):
Yeah, excellent. Yeah. Well, Dr. Grandin, I wanna wanna thank you for joining us today and to our loyal listeners. Thank you, you as well for joining us here. Once again, it's real science exchange. We hope you learned something. We hope you had some fun and we hope to see you next time here at the Real Science Exchange, where it's always happy hour and you're always among friends.
Commercial (41:45):
We'd love to hear your comments or ideas for topics and guests. So please reach out via email to anh.marketing at balchem.com with any suggestions and we'll work hard to add them to the schedule. Don't forget to leave a five star rating on your way out. You can request your Real Science Exchange t-shirt in just a few easy steps, just like or subscribe to the Real Science Exchange. And send us a screenshot along with your address and t-shirt size to anh.marketing at balchem.com. Balchem's Real Science lecture series of webinars continues with ruminant focused topics on the first Tuesday of every month. Monogastric focused topics on the second Tuesday of each month, and quarterly topics for the companion animal segment. Visit balchem.com/realscience to see the latest schedule and to register for upcoming webinars.