Google AI: Can you trust its advice about foods and health?
Want to know which foods are good for your memory, metabolism, or more? If you ask Google, the answers at the top of the page are from its “AI Overview.” Is Google’s artificial intelligence credible...or just popular? We checked AI’s sources. They may surprise you.
1. Check the source.
Start by clicking on the link symbol after each statement. Don’t be surprised if it’s a company that sells something that AI is recommending.
Your best bets are guidelines from health authorities like the National Institutes of Health, Centers for Disease Control and Prevention, and American Heart Association, especially if they cite evidence.
They’re better sources than websites like WebMD, MedicalNewsToday, or VeryWellHealth, which don’t always cite the strongest evidence.
And don’t be impressed by advice from hospitals, healthcare systems, or universities. They have experts on staff, but their websites may post articles written by staffers who don’t know the research, even if what they write has been reviewed by a health professional.
2. AI isn’t necessarily intelligent.
Google’s AI Overview gathers info from the internet. It doesn’t vet the information to see what’s based on the best evidence.
So if many websites make the same claim, it shows up on AI. In other words, it’s more a measure of what’s popular than what’s accurate. (We didn’t look at ChatGPT or Copilot.)
And websites copy each other. Why else would bell peppers, broccoli, and sweet potatoes show up on so many lists of foods that are rich in vitamin A or C? Plenty of other fruits and vegetables also fit the bill.
And nuts, fish, eggs, and dark chocolate appear on lots of lists for whatever ails you. They’re having a moment!
3. It’s a jungle out there.
Every website is competing for eyeballs. The more clicks they get, the more ad revenue, product sales, or subscribers they can get.
And there’s a built-in bias. If people want to know, say, what foods are good for skin, a website that offers a list of skin-friendly foods gets more clicks than one that says that foods won’t do a thing for your skin. Beware of the bias in favor of “this-food-is-good-for-whatever!” advice.
4. Don’t fall for the vitamin trap.
Yes, some foods are rich in vitamins you need. But unless you’re suffering from a true vitamin deficiency—which is extremely unlikely in the U.S.—don’t assume that those foods (or vitamins) can fight infections, protect your vision, prevent hair loss, etc.
And keep in mind that some studies can’t tell if a low vitamin intake causes, say, depression or memory loss...or if those problems cause people to eat fewer vitamin-rich foods.
5. AI’s answers keep changing.
Ask a question like “What foods are good for immunity?” on two different days—or on two different devices—and you’ll get somewhat different answers.
As Google says, “Generative AI is experimental.” Indeed.
Keep reading to see what we found on a few days in September and October.
What we asked Google AI
We googled 4 questions:
- What foods are good for memory?
- What foods are good for speeding up metabolism?
- What foods are good for preventing depression?
- What foods are good for preventing a urinary tract infection?
Google AI’s answers often came from websites that sell those foods or have no relevant expertise, like a language academy. What’s more, those answers were often contradicted by large clinical trials or other solid evidence. Unfortunately, the Internet is packed with unreliable advice about diet and health, and you can’t rely on Google AI to separate the good from the bad. Here’s a sampling of what we found.
Good for memory?
Foods may help, says AI. Its evidence? Not too reliable.
Eggs. Their vitamins B-6, B-12, and folic acid “may help delay cognitive decline.” Source: MaxHealthcare.in (a healthcare system in India).
But in multiple clinical trials, B vitamins had no impact on cognition.
Dark chocolate. Its flavonoids “may help with memory and learning.” Source: Spanish.academy (a website for learning Spanish).
Yet in the COSMOS trial, cocoa flavanols had no benefit for memory. (The daily dose was what you’d get in 600 calories’ worth of dark chocolate.)
Nuts. They improve memory recall. Source: Nuts ‘N More (an online nut store). No bias there!
True, in an industry-funded trial, 28 older adults who ate 2 ounces of mixed nuts a day for 16 weeks scored better on 2 out of 6 memory tests than when they ate no nuts. (On 1 of those 2 tests, though, they could recall only 1 more word after eating nuts.)
But in another industry-funded trial, on 657 older adults, those who were randomly assigned to eat 1 to 2 ounces a day of walnuts for 2 years scored no better on memory or overall cognition tests than those who ate none.
Oily fish. Its omega-3 fats “may help improve short-term and long-term memory.” Source: the Hindustan Times.
But a trial that randomly assigned 4,218 people to take either fish oil (1,000 mg a day) or a placebo for 2 to 3 years reported no difference in memory.
Bottom line
No foods have been proven to help memory. What about berries, which AI also lists? They’re part of the MIND diet, which has been linked to slower brain aging in studies that track people for years. MIND is a mix of a healthy Mediterranean diet and a DASH diet. But the only trial to test MIND on memory came up empty. Stay tuned.
Good for speeding up metabolism?
Want to burn more calories? AI has answers!
Coffee & green tea. “Caffeine in coffee can stimulate your metabolism and increase energy expenditure.” Source: MedicalNewsToday.com.
In the study cited by MedicalNews, people burned about 100 calories more per day after high doses of caffeine (270 to 1,600 mg, but not 150 mg). But neither caffeine nor green tea cause significant weight loss.
Bell peppers. Their lycopene “is good for your heart” and their fiber “can help lower cholesterol.” Huh? Source: IndiaTVnews.com.
Ginger. It “increases metabolism.” Source: Vitality4life.co.uk (which sells kitchen wares, fitness equipment, etc.). Yet in recent studies, ginger did squat for metabolism.
Eggs. “The body needs to use more energy to digest protein-based food like eggs.” Source: VMeals.ae (which sells meal plans in the UAE).
That may be true, but an egg has only 6 grams of protein, which supplies 24 calories. In theory, you’d use 20 to 30 percent of those calories to metabolize the protein, which works out to just 5 to 7 calories.
Water. “One study found that drinking 500 ml of water increased metabolic rate by 30%.” Source: MyJuniper.co.uk (a weight-loss website). But in later studies, water had little or no effect.
Nuts. “One study found that there was an elevation of energy expenditure after eating them.” Source: HollandandBarrett.com (which sells wellness and weight-loss supplements).
Yet according to a 2023 review funded by the nut industry, “There was no evidence to suggest that the long-term consumption of nuts affects [resting energy expenditure].”
Bottom line
Want to boost your metabolism? Get moving. The more you move, the more calories you burn.
Good for preventing depression?
Foods can help, says AI. Really?
Eggs. “The yolks contain B vitamins, which may help reduce depression symptoms.” Source: Piedmont Healthcare in Atlanta.
But in two large trials, vitamins B-6, B-9 (folic acid), and/or B-12 had no impact on the risk of depression. Nor did folic acid prevent depression in people with depressive symptoms.
Seafood. Fatty fish “contain omega-3 fatty acids, which may help with mood.” Source: WebMD.com.
But in a 5-year trial on 18,353 people, those who were randomly assigned to take fish oil (465 mg of EPA and 375 mg of DHA) every day had a 13 percent higher risk of depression than placebo takers.
Nor did fish oil prevent depression in people with depressive symptoms.
Bananas. Their vitamin B-6 “helps produce serotonin and dopamine, neurotransmitters that can help improve mood.” Source: MEDvidi.com (an online mental health center).
Yet in a large trial, taking 50 mg a day of B-6 for 7 years didn’t prevent depression. A banana has 0.4 mg.
Berries. They “may help decrease depression and anxiety levels.” Source: StraightTalkCounseling.org (a mental health clinic in California).
Yet in five small trials, dried blueberries (or juice or powder) either raised, curbed, or had no effect on depressive symptoms.
Dark chocolate. “A 2019 study found that people who ate dark chocolate had a lower chance of reporting depressive symptoms than those who didn’t.” Source: Deseret.com (a Utah news organization).
Did dark chocolate cause less depression? That study can’t tell.
Milk. It’s ”a good source of vitamin D, which may help ward off depression symptoms.” Source: VeryWellHealth.com.
Yet in a large trial, vitamin D didn’t prevent depression in people at average or high risk for depression.
Bottom line
To prevent or curb depression, try exercising for 75 to 150 minutes a week.
Good for preventing a UTI?
What foods can prevent a urinary tract infection? Just ask AI!
Cranberries and blueberries. They “contain plant compounds that prevent E. coli from attaching to urinary tract cells.” Source: A 2015 article from Urology Specialists of the Carolinas (urologyspecialistsnc.com).
But in 2020, the FDA found only “limited and inconsistent” evidence that cranberry products could prevent UTIs.
Foods high in vitamin C. It “makes urine more acidic, which may prevent bacteria from growing.” Source: Urology San Antonio (urologysanantonio.com).
Even huge doses of C (2,000 mg a day) don’t make urine more acidic.
Probiotics. Good bacteria in fermented foods like plain Greek yogurt, cheese, kefir, sauerkraut, kimchi, pickles, and apple cider vinegar “prevent bad bacteria from growing.” Source: urologysanantonio.com. Trials proving that those foods can prevent UTIs? Zip.
D-mannose. This sugar, which is found in “cranberries, apples, oranges, peaches, broccoli, and green beans...prevents bacteria from sticking to the lining of the urinary tract system.” Source: urologysanantonio.com.
But in clinical trials, D-mannose supplements didn’t prevent UTIs.
Cauliflower. It has “anti-inflammatory compounds that can help control inflammation throughout your body and your urinary tract.” Source: UGA Urology of Greater Atlanta (ugatl.com).
Trials to back that up? None.
Water. “Drink plenty of water.” Source: Hackensack Meridian Health, a health care network (hackensackmeridianhealth.org).
Yes! Women assigned to drink an extra 1½ liters of water a day for one year had fewer UTIs (1.7) than those who drank no extra water (3.2). (Some authors worked for Danone Research. Danone sells bottled water.)
Bottom line
Urologists are entitled to their professional opinions, but, water aside, the evidence for foods to ward off a UTI is skimpy at best.
Support CSPI today
As a nonprofit organization that takes no donations from industry or government, CSPI relies on the support of donors to continue our work in securing a safe, nutritious, and transparent food system. Every donation—no matter how small—helps CSPI continue improving food access, removing harmful additives, strengthening food safety, conducting and reviewing research, and reforming food labeling.
Please support CSPI today, and consider contributing monthly. Thank you.
More behind the headlines
5 food & supplement ideas to doubt
Weight and Health
6 food myths with bold claims and skimpy evidence
Fact vs. Fiction
Decoding the latest food & health headlines
Fact vs. Fiction
News you can lose
Fact vs. Fiction
How news can confuse: What's behind some recent headlines
Fact vs. Fiction