Podchaser Logo Rate on Podchaser

Episode 86

full
Published on:

30th Jun 2025

WELCOME TO OUR KITCHEN: We're talking about AI-generated recipes!

No doubt, the internet is being flooded with AI recipes. In fact, even unscrupulous publishing is being flooded by these generated recipes as small, independent publishers pop up who take advantage of the "free" content to package it as a book.

But there's some hope. AI recipes aren't all bad. And there are ways to identify them before you caught in the trap of making, oh, lemon meringue trout pie.

We're Bruce Weinstein & Mark Scarbrough, veteran cookbook authors who've written over three dozen cookbooks and who've been around the block several times already with AI content. We'll tell you how to spot an AI recipe, how to use AI to help in the kitchen, and what the dangers of AI recipes are.

We've also got a one-minute cooking tip about a handy kitchen appliance. And we'll tell you what's making us happy in food this week.

Here are the segments for this episode of COOKING WITH BRUCE & MARK:

[01:05] Our one-minute cooking tip: use a stick blender!

[03:11] How to spot, use, and stay away from AI-generated recipes.

[29:01] What’s making us happy in food this week: curried chili crisp and potato salad!

Transcript
mark:

Hey, I am Bruce Weinstein and this is the Podcast

2

:

Cooking with Bruce and Martin.

3

:

And I'm Marks Scarborough.

4

:

And together with Bruce, my husband,

we have written 37 cookbooks, including

5

:

Cold Canning, a book that is all about

making small batches of preserved and

6

:

fermented foods, and even laurs like

Triple Sec and even dessert sauces like

7

:

strawberry ice cream topping, small

batches, two or three jars that you

8

:

could put in the fridge or the freezer.

9

:

No pressure or steam.

10

:

Canner needed for anything, even

the fantastic Strawberry Preserves

11

:

and the Blackberry Cham in the book.

12

:

That's our latest cookbook,

but this is our podcast about

13

:

our passion, food and cooking.

14

:

We've got, as always, a

one minute cooking tip.

15

:

We're gonna talk all about.

16

:

AI generated recipes.

17

:

Maybe we won't cover everything

you need to know about them.

18

:

Mm-hmm.

19

:

Mm-hmm.

20

:

But we're gonna talk about

what we think about the coming

21

:

deluge of AI generated recipes.

22

:

Oh, I know.

23

:

And we'll tell you what's making

us happy in food this week.

24

:

So let's get started.

25

:

Bruce: Our one minute cooking tip.

26

:

You need to get yourself an immersion

blender sometimes called a stick blender.

27

:

I think

28

:

mark: it's like a stick

blender in the uk, right?

29

:

Mm-hmm.

30

:

And an immersion blender in the US mostly.

31

:

Oh, it's

32

:

Bruce: such a

33

:

mark: great

34

:

Bruce: tool.

35

:

These are just such great tools.

36

:

Imagine a rolling pin that has a

button on the top where you hold it,

37

:

and at the bottom there's a blade.

38

:

They can cut

39

:

mark: you to death.

40

:

They can slice

41

:

Bruce: you

42

:

mark: off.

43

:

Bruce: No, the blade is encased, right?

44

:

It's.

45

:

Was for safety, and when you push

the button, it spins so convenient.

46

:

If you are pureing a soup, you could put

the stick blender right in your pot of

47

:

soup, hit the button and puree that carrot

soup or that bean soup right in the pot.

48

:

Sometimes if I'm making jams and

it's a little too chunky, I could

49

:

put it in and just give it one or

two clicks and it pulverizes the

50

:

fruit to just the right consistency.

51

:

Right.

52

:

Of course, we're making mayonnaise.

53

:

It's just you can't beat it.

54

:

Right.

55

:

And it's a very.

56

:

Handy tool.

57

:

And

58

:

mark: it's also an easy way

to make smoothies, right?

59

:

Because if you wanna drink out of, not a

glass container, but a plastic container

60

:

that you make the smoothie in, you can

put the ingredients in there, stick

61

:

a stick blender in the bottom of it.

62

:

And it actually is easier cleanup,

I think, than a giant blender.

63

:

It's so easy because, and all of its

pieces, the giant blender has so many

64

:

pieces that need to get cleaned up.

65

:

Bruce: Yeah, the stick blender, you just.

66

:

Unscrew the blade part and

put it in the dishwasher.

67

:

Right.

68

:

That's how easy it is.

69

:

Right, right.

70

:

And usually they come with very long

cords so you can reach your stove and

71

:

your pot no matter where your plugs are.

72

:

They're really a great

73

:

mark: tool.

74

:

Yeah, it is a great way to

take your cooking over the top.

75

:

Okay, before we turn on to AI generated

recipes or turn off onto AI generators.

76

:

Piece, we wanna tell you that we

appreciate your support at this podcast.

77

:

It would be great if you could rate it

and if you could even write a review of

78

:

it that would be spectacular on the social

media platforms that you find this on.

79

:

Let's say you're on Apple

Podcasts or Spotify.

80

:

Give it a rating and if you can't even

stop and drop a review, even a short

81

:

review, like great podcast, that really

helps us because we prefer to remain

82

:

unsupported and will in fact remain

unsupported because we don't want.

83

:

To be beholden to any corporate overlords.

84

:

So I wanna say what we wanna

say, which is what we're about

85

:

to do about AI generated recipes.

86

:

So let's start talking about that.

87

:

Bruce: Okay.

88

:

I wanna start with the basic question.

89

:

What.

90

:

Is an AI generated recipe?

91

:

mark: Well, an AI generated

recipe is what you think.

92

:

It is a recipe that has been

generated by artificial intelligence.

93

:

It is essentially, let's say,

fed into chat, GPT or any of the

94

:

giant AI creators, and a recipe

is spit out for, I don't know.

95

:

Chicken thigh saute.

96

:

These recipes are made because the

technology that is underpinning the AI has

97

:

combed the internet, has gleaned all of

the recipes out there for chicken thighs

98

:

that are sitting out on the internet and

has, uh, taught, that's such a big word.

99

:

Mm-hmm.

100

:

But taught itself things about

chicken thigh sautes based on the

101

:

thousands, millions of recipes

that it can find online as it.

102

:

Combs through all the searches.

103

:

Bruce: But let me ask this

about AI generated recipes.

104

:

AI has no taste buds, no AI has

no idea of what things taste like.

105

:

No.

106

:

So AI is not using any sort of

human sense of what tastes good.

107

:

Well, to create these

108

:

mark: secondhand.

109

:

Here's the secondhand part.

110

:

It's very unlikely that you're gonna find

an AI generated recipe that uses, say.

111

:

A cup of dried thyme.

112

:

That's because having combed through

millions of recipes through searches,

113

:

the AI has discovered that most recipes

use a teaspoon, a half a teaspoon, a

114

:

teaspoon and a half of dried thyme.

115

:

And so it is now believing, I

know that's a personal human word.

116

:

Mm-hmm.

117

:

Believing, but it is.

118

:

It is using, its algorithmically

generated, and in fact its artificial

119

:

intelligence generated knowledge to

know that basically when you use dried

120

:

time, you use about a teaspoon of it.

121

:

So that's what it's going to do.

122

:

You're right, in terms of the fact

that a teaspoon of dried time may not

123

:

actually balance out what's in the pod.

124

:

Mm-hmm.

125

:

But it is learning it.

126

:

So here's what we're gonna do.

127

:

We're gonna talk about why AI

generated recipes are dangerous.

128

:

We're gonna talk about how you can spot

AI generated recipes on the internet.

129

:

But then we also, we don't wanna be

totally negative, so we wanna talk about

130

:

what's good about AI generated recipes and

there are good things to say about them.

131

:

It is the coming wave, no

matter what else we think.

132

:

So let's talk about.

133

:

The problems, and let's talk

about how to notice and be smart

134

:

about recipes you find online.

135

:

But let's also talk about

what can be good about this.

136

:

So let's start out why they're dangerous.

137

:

Okay.

138

:

I can tell you, I'll start out.

139

:

Okay.

140

:

One of the ways that they're dangerous

is that since AI is combing through not

141

:

only the internet, but books, since books

are being fed into AI generators too,

142

:

and the content is being scanned, the.

143

:

Cooking times the doneness levels.

144

:

How things are measured as cooked or

done is widely varied because there

145

:

are a lot of idiots online who think

a cake can be baked in 10 minutes.

146

:

There are a lot of people who think

that a pork roast is done in 20 minutes.

147

:

Remember that Barbara

148

:

Bruce: Kaka book roasting

from the eighties?

149

:

I do, and her idea was you take

a chicken, you put it in a 500.

150

:

50 degree oven for seven minutes,

151

:

mark: right?

152

:

It wasn't seven, but it was 20

and it was insane because almost

153

:

everything in that book was raw.

154

:

That book won a million awards.

155

:

So just say, but nothing worked.

156

:

Just, yeah.

157

:

So just say this problem of

weird recipe doneness and cooking

158

:

times and cooking weighs it.

159

:

It's been, it's been what?

160

:

Endemic, it's been a problem in

our industry for a long time.

161

:

But again, the.

162

:

AI's just slurping all of this up.

163

:

So it's gonna come out with, you

know, I don't know that you can do

164

:

a bone in chicken breast in perhaps

11 minutes on a grill, which is not

165

:

Bruce: possible.

166

:

Which is really weird because online there

does exist the USDA guidelines, right?

167

:

And international guidelines of what

temperatures meet needs to be cooked

168

:

at how long it needs to be held at

that temperature to be safe to eat.

169

:

And you would think that these ais.

170

:

Would be trained to look for

that information as well.

171

:

mark: You mean programmed?

172

:

Yeah.

173

:

Um, not trained.

174

:

It's not a dog.

175

:

Right.

176

:

Um, you mean programmed because that would

take care of that problem, wouldn't it?

177

:

This is the problem with AI

generation right now is that

178

:

all information is value flat.

179

:

So it's going across, it's scanning

across not only your USD guidelines,

180

:

but dozens and hundreds and thousands

and millions of websites and books

181

:

and recipes, and it's treating.

182

:

All that information as

if it was all value flat.

183

:

Mm.

184

:

So it's all exactly of the same weight.

185

:

And when it does that, it looks across it

and it says, oh look, the majority of, I'm

186

:

gonna go back to my chicken breast recipe.

187

:

The majority of chicken breasts on

the grill online get cooked between,

188

:

I don't know, 12 and 15 minutes.

189

:

So therefore, mm-hmm.

190

:

That must.

191

:

Be the appropriate cooking

time for a chicken breast.

192

:

The problem is chicken bony and chicken

breast can run anywhere from what,

193

:

half a pound to over a pound, right?

194

:

Bruce: And it should be cooked to

165 degrees as all poultry should,

195

:

according to the USDA, and that is

rarely taken into consideration by ai.

196

:

It doesn't look at those kind of

safety numbers, so you're missing.

197

:

Those safety protocols.

198

:

mark: Yes, you are.

199

:

And you're also missing basic

safety protocol techniques

200

:

like we write in our recipes.

201

:

So let me just say like about

cold canning, we made sure that in

202

:

every recipe in cold canning it,

it specifically says to ladle the

203

:

blackberry jam or the blueberry chutney

or the barbecue sauce into clean.

204

:

Jars.

205

:

That's a safety protocol we've put

into the recipe and we explain in the

206

:

introduction, you can run your jars

through a hot cycle, a dishwasher,

207

:

you can run them in hot water, you

can boil them on the stove if you

208

:

want to get really crazy, but you

can do all of this to clean the jars.

209

:

That's a safety protocol that a writer

has written into a recipe, and AI probably

210

:

doesn't see that as a safety protocol.

211

:

Bruce: No, and that's the

nice thing about a book.

212

:

Is you're gonna have an

introduction to the book.

213

:

You're gonna have an introduction

to each chapter, and you're gonna

214

:

have a head note, which all adds up

together to give you an understanding

215

:

of what needs to happen in that

recipe, both to make it delicious,

216

:

successful, and also to make it safe.

217

:

mark: Right?

218

:

And let, let.

219

:

There's also the problem with AI that

there is often poor ingredient handling.

220

:

For example, uh, I found several as I

was researching the, for this episode,

221

:

I found several recipes online,

clearly AI generated, where it talked

222

:

about just chopping up scallions.

223

:

Well, first of all, you don't.

224

:

Chop scallions, you slice them.

225

:

And secondly, any good recipe would

tell you to cut off the wagley roots at

226

:

the end of it and cut off any squishy

green parts at the top, and maybe

227

:

say the white only or the green part.

228

:

Only, but it doesn't know any of that.

229

:

Mm-hmm.

230

:

No, again, no is such a human word.

231

:

It doesn't recognize any of that.

232

:

And so it just says, chop up the scallion.

233

:

Well, you might say, I know to

take the roots off scallions, but a

234

:

lot of people might not know that.

235

:

Oh

236

:

Bruce: my goodness.

237

:

No.

238

:

Let's go back to that famed thing

where we have to put drain the

239

:

pasta in a colander set in the sink.

240

:

Yeah.

241

:

And you think that that is, duh.

242

:

We have to put that in our

books because people don't know.

243

:

mark: Yes, that's right.

244

:

There are actually readers who say,

I drained the pasta on my counter and

245

:

the water went all over the kitchen.

246

:

And I know you think,

my gosh, are you crazy?

247

:

But in fact, it's the truth and

not to make fun of those people.

248

:

Because if you don't know how to

cook, you don't know how to cook.

249

:

Exactly.

250

:

You don't know what

251

:

Bruce: to do.

252

:

No.

253

:

If you never baked a cake and it says

butter a cake pan, how do you know you're

254

:

only supposed to butter the inside?

255

:

If you never bake the cake, how would

you know that's another, so now we say.

256

:

Grease or butter the

257

:

mark: inside of a pan.

258

:

That's another infamous one that actually

happened to us early in our career

259

:

that we had someone write in and say

they grease the pan and then it slipped

260

:

out of their hands because it was so

greasy and it was like, oh my gosh, I

261

:

didn't say grease the inside of the pan.

262

:

So this person clearly

greased the whole pan.

263

:

But again, not to make fun of this person

because if you don't know, you don't know.

264

:

If you don't know, it's not

like cooking is intuitive.

265

:

Okay, so let's, let's move on

and talk about how do you spot an

266

:

AI recipe because they are now.

267

:

Absolutely everywhere across

the internet space, my Facebook

268

:

feed is filled with them.

269

:

And I want to give you some,

um, ways that I've researched

270

:

this and thought about this.

271

:

I spent the last few days looking through

all those recipes and uh, I can tell

272

:

you how I spotted the ones that are ai.

273

:

And the first I can tell you is that

there are unusual ingredient combinations,

274

:

things that you would never do.

275

:

Mm-hmm.

276

:

I actually found a recipe

for a raspberry shrimp.

277

:

Stir fry.

278

:

Bruce: Okay.

279

:

Now while that sounds pretty gross,

uh, disgusting, I do need to say that

280

:

unusual ingredient combinations are kind

of what I get paid to come up with Yes.

281

:

In my career.

282

:

That's right.

283

:

So it is an interesting dilemma.

284

:

Here we are.

285

:

Here.

286

:

I am a trained chef.

287

:

And I am the chef in our duo.

288

:

And when it comes to creating the recipes

that Mark and I talk about for our books,

289

:

I have to come up with really interesting

and unusual flavor combinations.

290

:

I don't think I would put raspberries and

shrimp together, but I might, no, it's

291

:

the stir fry part, but I might do a, I can

292

:

mark: actually imagine raspberries

and shrimp, I guess, but.

293

:

But IS are shrimp in a stir

294

:

Bruce: fry.

295

:

Yeah, well, like orange and

shrimp in a stir fry work.

296

:

Then why wouldn't an AI think

any fruit is gonna work?

297

:

Of course it would.

298

:

So it, it's taking its cues

from people like us who create

299

:

recipes, but it has no finesse.

300

:

mark: It's true.

301

:

And this isn't the only way to

recognize AI generator recipes.

302

:

It's one of what we're gonna say.

303

:

Alright.

304

:

Eight ways to recognize them.

305

:

So you mean to look out for really

strange combinations of things

306

:

that you think, wait, there's

no shrimp and a chocolate cake?

307

:

Or we,

308

:

Bruce: I'm thinking, no, you're

309

:

mark: right.

310

:

No, there is not.

311

:

No, there is not.

312

:

Or I don't know.

313

:

There's no.

314

:

Baking powder flavored muffins, those

do not exist or should not exist.

315

:

So again, unusual ingredient

combinations is typical and

316

:

also nonsensical ingredients.

317

:

I actually found a recipe online for, hey.

318

:

Pudding And I don't, and this didn't

even talk about, I mean, maybe this is a

319

:

thing that, that, you know, I don't know.

320

:

In the middle ages, people made some kind

of egg pudding out of hay, but this didn't

321

:

include any instructions on cleaning hay

or what you would need to know about hay.

322

:

It's just like taking your yard

clippings and making a pudding out of

323

:

Bruce: it.

324

:

That's an interesting example

'cause a few years ago we went.

325

:

Through some Nordic cookbook that

passed our desk and in it was a recipe

326

:

for hay ice cream, and this chef had

taken some hay, I don't know if it was

327

:

cleaned or whatever, and infused it

into warm cream to get that hay flavor

328

:

into the cream, then strained it and

used that cream to make ice cream.

329

:

All the AI needs is to see that once

which is real, and start saying,

330

:

oh, we can make hay pudding, we can

make, you know, hay, anything hay

331

:

bread, and it doesn't make sense.

332

:

It made sense for that one chef.

333

:

But that was it.

334

:

mark: Yeah.

335

:

I, I think that, again, you can find

counter examples to all of these,

336

:

but I think the point is to really

watch out for nonsensical things.

337

:

Mm-hmm.

338

:

In the ingredients.

339

:

And also I found a recipe in for, on

online and doing research for this

340

:

podcast that was for lentil bread.

341

:

Okay, great.

342

:

You know, I mean, it is a thing.

343

:

Sure.

344

:

You can make it with lentil flour.

345

:

I'm sure you could make bread with

cooked lentils and, oh God, that sounds

346

:

Bruce: really good.

347

:

Actually, I'm,

348

:

mark: I'm sure that's a.

349

:

Thing.

350

:

However, the, the photo for

the recipe was a slice of what

351

:

looked like White Wonder bread.

352

:

Mm-hmm.

353

:

It was white US sandwich bread.

354

:

That recipe photo does not match.

355

:

Mm-hmm.

356

:

What the recipe promises.

357

:

So this is an immediate

flag that this is an.

358

:

AI recipe.

359

:

And may I say also in this case,

with the lentil bread, there was

360

:

no instruction of cooking it.

361

:

You heated your oven, but it

never said to put it in the oven.

362

:

You just mix the batter

and pour it into a pan.

363

:

And then it said cool for five minutes.

364

:

Mm-hmm.

365

:

So what does that mean?

366

:

Right?

367

:

Yeah.

368

:

No, that, that's clearly

an AI generated recipe.

369

:

But again, the picture didn't

match the recipe at all.

370

:

And.

371

:

That that's a problem in

Codebooks too, but it is.

372

:

You should really watch

it out for it online.

373

:

Bruce: It's a big problem in bloggers

and websites, which now I'm thinking

374

:

maybe all these years that we've

seen pictures that don't match.

375

:

It was the beginning of AI and

people just trying to fake you out.

376

:

Yeah.

377

:

mark: Yeah.

378

:

I also think that it's

part of the overall.

379

:

Fake out that occurs in the food industry.

380

:

I mean, Bruce and I fought this

forever when we were writing all our

381

:

instant pop books, because of course

there are a lot of very popular

382

:

instant pop books and I'm not gonna

name any of the big ones right now.

383

:

Ours.

384

:

Bruce: Instant Pop Bible.

385

:

Well, yes,

386

:

mark: of course.

387

:

Yes, we did really.

388

:

Well with the instant Bible, but there

were some that sold in the millions of

389

:

copies and they would always pull, I don't

know, filet mignons outta the Instant Pot.

390

:

And they were perfectly browned.

391

:

And I'm sorry, you cannot

pressure cook something to brown.

392

:

You just cannot

393

:

Bruce: prime ribs.

394

:

That came out like as if I'd had them in

the oven at 3 75 for four hours, right?

395

:

They were beautiful.

396

:

No.

397

:

Mm-hmm.

398

:

Doesn't work.

399

:

No,

400

:

mark: it doesn't work.

401

:

Okay.

402

:

So another way is vague

or missing measurements.

403

:

AI is not good.

404

:

Yeah, it might become good in the future,

but it's not good yet at recognizing

405

:

accuracy in ingredient measurements.

406

:

So it's gonna use terms like sum or

a bit of, and I know you're gonna

407

:

say this is the way people cook.

408

:

They throw sum of or a bit of,

but when you have, it's a pinch

409

:

Bruce: of, and that,

410

:

mark: right.

411

:

But when you have an ingredient amount.

412

:

In the ingredient list, one

teaspoon dry time, and then it

413

:

says, use a bit of the time.

414

:

Bruce: No.

415

:

Use the time that, that makes no sense.

416

:

No use the time,

417

:

mark: right.

418

:

What do I do with the rest of it then?

419

:

Okay, so watch out for those

vague and missing measurements.

420

:

The

421

:

Bruce: other thing is I often

find in AI developed recipes

422

:

that the cooking instructions are

just overly simplistic, right?

423

:

They are.

424

:

There's no finesse and there's no detail.

425

:

It'll say cook it until done.

426

:

What does that mean?

427

:

Right?

428

:

What does that mean for a bread?

429

:

What does that mean for a muffin?

430

:

What does that mean for a burger?

431

:

What does that mean for a steak?

432

:

What does that mean for a chicken?

433

:

What does that mean for a pudding?

434

:

Each one of those things has a

different way to tell whether it's

435

:

done, whether the toothpick is clean,

whether the meat's at 140 degrees.

436

:

Yep.

437

:

Whether the pudding is

set or still jiggly.

438

:

There are no details like that

in so many AI generated recipes.

439

:

mark: So we've given you five ways.

440

:

We're gonna go into six,

but lemme just review.

441

:

So unusual ingredient combinations,

nonsensical ingredients, photos

442

:

that don't match the recipe.

443

:

Vague or missing measurements.

444

:

Overly simplistic instructions.

445

:

And then here's a sixth one.

446

:

And that is, um, related to me

the writer in our team, which is

447

:

unusual formatting regularities.

448

:

Lemme explain what I mean by that.

449

:

That is when you look at the recipe.

450

:

Every single paragraph of the

method of how you make this

451

:

thing is exactly the same length.

452

:

So the generator is in, in a sense,

creating paragraphs that look alike.

453

:

Nobody's recipe actually always has the

same length of each step, and you'll

454

:

also notice in a lot of AI generated

recipes that the sentence structure is.

455

:

Invariable it's subject verb objects,

you know, put the time in the

456

:

bowl, put the flower in the bowl.

457

:

It's, it's absolutely mechanical.

458

:

And every single sentence seems

exactly like every other sentence.

459

:

This is a way you know that it

has been unusually formatted.

460

:

And it could be an AI recipe, I suppose

there are people who write like that.

461

:

Well,

462

:

Bruce: they're not gonna

sell many books then.

463

:

mark: Well, and, but this is when.

464

:

Specifically talking

about the internet, right?

465

:

Yep.

466

:

So I suppose there are bloggers who write

like that, who never got beyond fifth

467

:

grade English, but most people vary the

length of their paragraphs and they alter

468

:

the structure of their sentences because

it just gets so boring to read the same

469

:

sentence over and over and over again.

470

:

Same sentence format.

471

:

Bruce: And that leads into the next thing,

which is unlikely or non-existent authors.

472

:

Mark just said, if you have something

that seems very mechanical, it probably

473

:

wasn't written by a person, right?

474

:

So an AI recipe may be attributed

to a blog or even a cookbook, but

475

:

chances are that's a fake author.

476

:

They're fake author

bios, fake author photos.

477

:

Yep.

478

:

You know, look for red flags that are like

a food blog being described as a cookbook.

479

:

Which what?

480

:

That makes no sense, right.

481

:

You know, with a list of recipe

titles in the bio or there's a

482

:

lack of verifiable online presence

or social media accounts for that

483

:

mark: author.

484

:

Okay.

485

:

I can give an example for this that

is actually from our life right now.

486

:

Bruce made a dessert for a dinner

party, uh, just last night as we're

487

:

recording this and it's a dessert.

488

:

Chinese dessert where you make from

scratch a very, very soft tofu,

489

:

which you did, you, you coagulated.

490

:

It's a

491

:

Bruce: silken style tofu,

492

:

mark: right?

493

:

You coagulated soy milk with

this coagulant that you use.

494

:

Mm-hmm.

495

:

And you make a very soft, uh, tofu.

496

:

And then you pour in

your case, what was it?

497

:

A ginger syrup?

498

:

Bruce: Ginger, orange, and omanis syrup.

499

:

Okay.

500

:

mark: Wow.

501

:

Okay.

502

:

You pour that over the top and

then you put little threads of.

503

:

Uh, of orange zest on top of

all of that, and he served it.

504

:

And so we were looking up recipes for

this, and there were tons of recipes,

505

:

as you can imagine, online for this

very classic Sichuan style dessert.

506

:

And it's not really dessert

in Chinese cuisine, but sweet.

507

:

But we were serving him as a dessert.

508

:

Okay.

509

:

Some of the recipes that we found.

510

:

Were on food blogs, and on at least two

occasions I would look at the recipe

511

:

and then I would look at the, you know,

smiling picture of the person mm-hmm.

512

:

At the top of the thing.

513

:

So I googled the person in each

case, and there was absolutely

514

:

no other presence for that person

except that blog, which means Right.

515

:

That, that's an AI generated blog designed

to sell the advertisements on the page.

516

:

Bruce: Right.

517

:

So what I did, just to continue

on and show a little bit of how I.

518

:

Overcame that.

519

:

I found a website on the Omnivores

Cookbook website where this woman,

520

:

Maggie Ju, who runs that and has written

books, tested making this recipe with

521

:

five different kinds of coagulants.

522

:

She used two different techniques

for each kind of coagulant, and

523

:

I actually had a question about

what I was doing and post it.

524

:

On her website and she responded within

five hours with her idea of what I needed

525

:

to do if I was tripling the recipe.

526

:

There you go.

527

:

And it was great and it worked.

528

:

So you have to be sort of

interactive and see if they're

529

:

responding and are they real?

530

:

mark: That's right.

531

:

Uh, you probably know if you've listened

to this podcast that several years ago

532

:

I got very heavy into vegan cooking

and making this kind of really wild

533

:

new vegan cuisine, which is not like

when I was a kid and vegan cuisine met.

534

:

Icky tofu and brown

535

:

Bruce: rice, steam broccoli, and

if you are lucky, sesame sauce.

536

:

Oh god, cold ses sauce, cold, pour over a

537

:

mark: limp, wilted greens.

538

:

It's just disgusting.

539

:

So, um, you know, I, I wanted

to get away from the moose

540

:

wood stuff, sorry, moose wood.

541

:

But I wanted to get away from that and

I wanted to find the modern vegan stuff.

542

:

So I did, and I found a

lot of plaques online.

543

:

A lot of the recipes I started

trying were all failures.

544

:

And then it occurred to me, wait

a minute, these aren't really

545

:

blogs by legitimate food bloggers.

546

:

These are instead AI generated blogs.

547

:

And then I discovered, yes, of course

I have to go out to social media and

548

:

see if these people actually exist.

549

:

Do they, are they, you know, are they.

550

:

Pumping their blog up on Facebook or

Instagram or TikTok, and then I could

551

:

be a little more assured of the recipes.

552

:

So again, and the unlikely or

non-existent author is a key way.

553

:

You know, this is an AI generated

recipe, or even AI generated food blog.

554

:

And finally look at the comments.

555

:

Bruce: Mm-hmm.

556

:

mark: They reviews if they're

very bot like it's probably ai

557

:

if every comment is terrific.

558

:

Great.

559

:

Loved it.

560

:

This was good.

561

:

Yeah, loved it.

562

:

My kids loved it.

563

:

Like with no personality behind it.

564

:

Then you probably know with

repetitive and bot like comments

565

:

that this is an AI generated.

566

:

Recipe and or blog

567

:

Bruce: online recipe blog.

568

:

And are there responses

to those comments too?

569

:

And what are those responses like?

570

:

Is every one of them thanks?

571

:

Well, you know, that's

probably not a real person.

572

:

mark: Oh yeah.

573

:

Or even responses at all, because a lot of

the programmers don't seem to know that.

574

:

In fact, if you set up this blog

with all these recipes on it mm-hmm.

575

:

That you need to actually have a re.

576

:

Bonds, any, any, uh, food

writer would know to respond to

577

:

people who comment on things.

578

:

Right?

579

:

So it, it's all part of it.

580

:

And you might say, why are all of

these AI blog and food blogs happening?

581

:

And part of it is, as I said

earlier, it's to sell advertising.

582

:

Yep.

583

:

Bruce: That's what it's all about.

584

:

It's

585

:

mark: so you land on the page, you see the

recipe there, but what you really see, or.

586

:

All of the incessant ads around it,

and they're being generated in order

587

:

to be essentially freeway billboards.

588

:

Mm-hmm.

589

:

Um, and the content in the, in the center

of it is irrelevant to all the ads that

590

:

are running all around it, and even in

the pop-up ads that are coming off of it.

591

:

That's why they're being created partly.

592

:

And um, this is a problem.

593

:

Okay.

594

:

So there's are ways to spot AI recipes.

595

:

There's why we think they're dangerous.

596

:

Now let's talk for just a minute.

597

:

Mm-hmm.

598

:

About why we think they are good.

599

:

Because there are

reasons why they're good.

600

:

Bruce: There are real reasons.

601

:

I love using AI generated

recipes for ideas.

602

:

I can go to buzzfeed and they can,

they have an AI recipe generator.

603

:

I could put in things that I want to

cook and it will throw ideas at me.

604

:

Right.

605

:

And I love that because

it's sometimes if I'm in a.

606

:

Funk and I can't think of something new,

and I know this is what I do for a living,

607

:

but sometimes my brain gets really tired

and I can't think of a new chicken dish.

608

:

Well, an AI can help me figure that out.

609

:

Well, what should I make for dinner?

610

:

I know I have bacon, I have

chicken thighs, I have rosemary,

611

:

and I have olives right now.

612

:

Yes, I can come up with five things

right off the top of my head.

613

:

I know.

614

:

I know how to cook.

615

:

But let's say you have all those things

and you're not an experienced cook.

616

:

Tell Buzzfeed's AI generator,

that's what you have.

617

:

It'll spit out a beautiful recipe for you.

618

:

Well, e, even if

619

:

mark: you go into Google search at

this point and you put in how to

620

:

pan sear chicken thighs, it's gonna

first give you the AI response.

621

:

Mm-hmm.

622

:

And actually, if you look down that AI

response, I know it's easy to dismiss

623

:

it, but if you look down at, if you

know what you are doing, it might

624

:

spur you on to do something different.

625

:

Mm-hmm.

626

:

But I think.

627

:

In all of what we're saying is what's

good about AI generated food content

628

:

and even AI generated recipes,

what can be good about them is,

629

:

uh, they can spur your creativity.

630

:

Yep.

631

:

Presuming that you already

know how to cook, right.

632

:

Bruce: If you don't have a

clue as how to cook a chicken

633

:

thigh, if you don't know how to.

634

:

Bone, a chicken thigh.

635

:

That's right.

636

:

You don't even know how to take

the skin off a chicken thigh.

637

:

That's right.

638

:

Then those are not

gonna be helpful to you.

639

:

But if you are a pretty good cook or

even just an everyday cook and you

640

:

know how to put dinner on the table,

then an AI generated response to I

641

:

have these things in the house can,

as Mark said, spur your creativity

642

:

and help you make something really

643

:

mark: interesting.

644

:

That's right, and I think that

that's, if you're, if you're an

645

:

established cook and you know what

you're doing, this can be a great.

646

:

Tool to help you in fact do

something that is creative and

647

:

interesting in the kitchen.

648

:

If you're starting out or you're

trying to, as I was, let's say,

649

:

trying to explore modern vegan

cuisine, it's actually a detriment.

650

:

It's actually a, a hindrance to

what you're trying to do because

651

:

you're gonna end up with failures.

652

:

And I ended up with a million failures.

653

:

Oh, I should say one more thing about my.

654

:

Failure list just before we pass

on to the end of the podcast.

655

:

Oh, what, what was it?

656

:

Six years ago, seven years ago, I

got totally into classic British

657

:

desserts and classic British sweets.

658

:

Do you remember this?

659

:

Oh, boy.

660

:

And I, I, I parkin story parkin, and

you can go to our YouTube channel

661

:

and still see my video from like

seven years ago of making parkin.

662

:

If you don't know what that is, it.

663

:

We'll help you survive the

nuclear holocaust because it's

664

:

Bruce: a lard based pastry dessert

grid, I don't know what you call it,

665

:

mark: and you have to ripen it for weeks,

and then it lasts like indefinitely.

666

:

So it's nothing like

667

:

Bruce: ripened lard and wheat products.

668

:

mark: Yes.

669

:

Nothing like it.

670

:

So I.

671

:

I was investigating all these

things, and a British friend of

672

:

mine kept saying to me, try to find

a good recipe for a lardy cake.

673

:

If you're from the uk, you know

exactly what I'm talking about.

674

:

If you're from the US you have

no idea what I'm talking about.

675

:

But she was like, I remember I was a kid.

676

:

My grandmother made lardy cake, and I've

never been able to find a good recipe.

677

:

I went through dozens and dozens

of food blogs, making, and I

678

:

made all of their hard cakes.

679

:

You, the Lord he used and

they were all garbage.

680

:

Mm-hmm.

681

:

They were all garbage.

682

:

I could never make that thing

work, no matter how hard I tried,

683

:

based on the recipes I found.

684

:

Online, and that was pre ai.

685

:

God only knows what lar cake recipes

look like now with AI because it's

686

:

taking all of those rancid recipes from

people thinking they're making lar cake.

687

:

Remember that one that I made that you.

688

:

Baked it in a glass bowl and at the end

I pulled it outta the oven and it was

689

:

just this lump of dough sizzling in lard.

690

:

It was, that had been

released from the dough.

691

:

Bruce: It was a dough

ball, right in lard sauce

692

:

mark: sizzling in hot lard.

693

:

Bruce: Nice.

694

:

mark: It was disgusting.

695

:

So again, if you know what you're doing

and I didn't, with L cake, you can

696

:

actually come up with something decent

with an AI recipe generator or even

697

:

the Google recipe generator that occurs

if you do it in the search engine.

698

:

Okay.

699

:

That's all we have to say about.

700

:

AI and AI recipes will, I'm sure have

more to say in future episodes 'cause

701

:

this is an ever-changing landscape.

702

:

But let's just say for a moment

that we're certainly glad that

703

:

you're a part of this podcast.

704

:

And if you're interested in

acquiring our book called Canning,

705

:

look in either the player for this

podcast or look on our website.

706

:

And there are ways that

you can order that book.

707

:

Now there is in fact an order link

right in the player for this podcast.

708

:

So if you're interested in this

small batch canning idea without

709

:

a lot of work and making just a

couple choices, something, check

710

:

it out and you can get it there.

711

:

Okay, let's go on to the last and

traditional, uh, part of this podcast.

712

:

What's making us happy in food this week?

713

:

Bruce: It's something else

from last night's dinner party.

714

:

Besides that soft tofu I have in the past

said that Sichuan meat pies have made

715

:

me happy and I did make them last night.

716

:

And yes, they made me happy.

717

:

But what made me happier was a recipe

from our new book, cold Canning and

718

:

Curried Chili Crisp, right that I

served alongside those meat pies.

719

:

Now, this Curry Chili Crisp was one of

the hottest spiciest recipes in the book.

720

:

And I had, it's odd, I was a little

afraid there were some new people who

721

:

had never eaten at our house before

that were at this dinner party.

722

:

So I had the chili crisp and

little bowls on the plates.

723

:

You could spoon it in every single person.

724

:

Ate every drop of chili crab and it will

dipping and slathering it onto their meat

725

:

mark: bucket.

726

:

That stuff was not for the faint of part.

727

:

I mean, I've had, Bruce has been

sitting in the house for a while now.

728

:

Bruce made it a few weeks ago.

729

:

Uh, we actually posted a video,

didn't we, of it on TikTok, don't I?

730

:

If not, I

731

:

Bruce: will.

732

:

'cause that's time

733

:

mark: to make some more.

734

:

I, I think so.

735

:

Anyway, um, it, it's really hot.

736

:

It's incredibly burny, but it is

incredibly delicious and I ate all mine.

737

:

Mm-hmm.

738

:

I ate every bit of that chili

crisp on the ses meat guys want.

739

:

So I guess what's making me happy

in food this week is a summer

740

:

treat, and that is potato salad.

741

:

And this week Bruce was grilling

chicken thighs because it's this summer

742

:

and also this week in New England

it was 5 billion degrees Fahrenheit.

743

:

So that meant you grilled outside

and he made a potato salad and.

744

:

Uh, I have to say that, uh, we

don't put hard boiled eggs No.

745

:

In our potato salad.

746

:

Gross.

747

:

But he put raw broccoli florets

in there and it was delicious.

748

:

Mm-hmm.

749

:

Mm-hmm.

750

:

I had never had, I know it's such

a silly, simple thing, but I never

751

:

had bro raw broccoli in a potato

salad, and it just was fantastic.

752

:

I ate way more than my fair share.

753

:

That was good.

754

:

And you made a ton of it, so we

had it over a couple of meals.

755

:

Mm-hmm.

756

:

Hmm.

757

:

I don't know.

758

:

Potato salad is just summer.

759

:

I think really for me,

it's, uh, especially from

760

:

where I'm from, it's summer.

761

:

If you set the mayonnaise potato

salad out on the back deck for

762

:

like three hours and then eat it.

763

:

Now it's summer.

764

:

Mm.

765

:

What is summer without,

uh, bad potato salad.

766

:

Okay.

767

:

Gross.

768

:

Gross.

769

:

Um, anyway.

770

:

We didn't, we refrigerated

ours and it was fine.

771

:

So raw broccoli and data

salad is an amazing thing.

772

:

So that's what's making us

happy in food this week.

773

:

Thanks for joining us on this podcast.

774

:

Thanks for being a part of this journey.

775

:

We appreciate your time spent

with us, and we hope you learn

776

:

something about AI generated recipes.

777

:

Bruce: And I wanna remind you all

that we have a Facebook group called.

778

:

Cooking with Bruce and Mark.

779

:

You can go share what you're eating

there with us, and also we have a TikTok

780

:

channel cooking with Bruce and Mark.

781

:

Go to TikTok, check out our channel.

782

:

We have videos of us cooking.

783

:

We have videos of us talking

about our life and what we like

784

:

to eat and don't like to eat.

785

:

So please go to our TikTok

channel and subscribe there to

786

:

cooking with Bruce and Mark.

787

:

And don't forget to come back for another

episode every week Cooking with Mark.

All Episodes Previous Episode

Listen for free

Show artwork for Cooking with Bruce and Mark

About the Podcast

Cooking with Bruce and Mark
Fantastic recipes, culinary science, a little judgment, hysterical banter, love and laughs--you know, life.
Join us, Bruce Weinstein and Mark Scarbrough, for weekly episodes all about food, cooking, recipes, and maybe a little marital strife on air. After writing thirty-six cookbooks, we've got countless opinions and ideas on ingredients, recipes, the nature of the cookbook-writing business, and much more. If you've got a passion for food, we also hope to up your game once and a while and to make you laugh most of the time. Come along for the ride! There's plenty of room!

About your host

Profile picture for Mark Scarbrough

Mark Scarbrough

Former lit professor, current cookbook writer, creator of two podcasts, writer of thirty-five (and counting) cookbooks, author of one memoir (coming soon!), married to a chef (my cookbook co-writer, Bruce Weinstein), and with him, the owner of two collies, all in a very rural spot in New England. My life's full and I'm up for more challenges!