Latest / Challenges of Faith Radio Program / The 3 different AI stories! Ep 3458

Transcript

The 3 different AI stories! Ep 3458 · Challenges of Faith Radio Program

• 0:00
  Speaker 1: Welcome back to Challenges of Faith radio program. I'm Gary
• 0:03
  Speaker 1: McKenzie on episode 3458. AI, AI, AI. Well, I know
• 0:14
  Speaker 1: you've heard that AI now predicts your age. And if
• 0:18
  Speaker 1: you have survived cancer. And all it takes is a
• 0:22
  Speaker 1: face scan. A new artificial intelligence model can measure someone's
• 0:28
  Speaker 1: biological age and predict their chances of surviving cancer just
• 0:34
  Speaker 1: by looking at an image of the person's face. Researchers
• 0:39
  Speaker 1: say scientists at Mass General Brigham Hospital in Boston built
• 0:46
  Speaker 1: a new AI model that can provide what they're calling
• 0:49
  Speaker 1: a non-invasive window into health. According to a published study
• 0:55
  Speaker 1: released in the Medical Journal, the Lancelet Digital Health, by
• 1:00
  Speaker 1: simply scanning a selfie, the AI can determine how a
• 1:04
  Speaker 1: person's cells, tissues, and organ systems are functioning as well
• 1:09
  Speaker 1: as his or her likelihood of surviving cancer. Facial appearance
• 1:15
  Speaker 1: offers a non-invasive window into health, the study states. We
• 1:21
  Speaker 1: built Foundation Artificial Intelligence for Health Recognition Face, which is
• 1:27
  Speaker 1: known as FAHR. A foundation model trained more than 40
• 1:33
  Speaker 1: million facial photographs and fine-tuned it for two distinct tasks,
• 1:39
  Speaker 1: biological age estimation and survival predictions. The study's researchers took
• 1:47
  Speaker 1: these more than 40 million unlabeled photographs from Web Face 42M,
• 1:52
  Speaker 1: a data set comprised of 42 million facial recognition images
• 1:57
  Speaker 1: representing 2 million identities taken from across the Internet. According
• 2:04
  Speaker 1: to AI research discovery platform Emergent Mind, a research team
• 2:11
  Speaker 1: In Beijing, Imperial College, London, and China-based X-Forward AI developed
• 2:18
  Speaker 1: a data set in collaboration with each other. After running
• 2:23
  Speaker 1: these data sets through the FAR models, researchers fine-tuned the
• 2:30
  Speaker 1: AI on 34,363 photographs of patients with cancer from a
• 2:34
  Speaker 1: research center in the Netherlands to detect signs of mortality risk,
• 2:40
  Speaker 1: the study said. The models outperform a previous age estimation
• 2:45
  Speaker 1: model in survival procrastination. This is the report found, and
• 2:51
  Speaker 1: to our knowledge, this is the first report of an
• 2:55
  Speaker 1: AI algorithm that can directly predict cancer patient survival from
• 3:00
  Speaker 1: a simple face photograph. The study in the Lancelot Digital
• 3:05
  Speaker 1: Health said, among 122 cancer patients found to have a
• 3:11
  Speaker 1: high risk of death More than half, 52.5%, died within
• 3:16
  Speaker 1: 90 days of the assessment, as predicted. According to the study,
• 3:23
  Speaker 1: the algorithm assesses age by looking at the area around
• 3:26
  Speaker 1: the temples, the forehead above the eyes, and the line
• 3:30
  Speaker 1: that connects the nose down to the corner of the mouth. Dr.
• 3:35
  Speaker 1: Raymond Mock, one of the study's authors, explained to CNN
• 3:42
  Speaker 1: that which... has been discussed. I'm going to switch topics now.
• 3:49
  Speaker 1: According to Cyber News, ChatGPT told laid-off Amazon engineer to
• 3:58
  Speaker 1: stop taking life-saving drugs and lie to doctors, says by
• 4:04
  Speaker 1: Nuxios AI, reviewed by Cyber News staff. OpenAI is being
• 4:13
  Speaker 1: sued following the death of a 31-year-old Amazon software engineer
• 4:18
  Speaker 1: who took medical advice from ChatGPT. Bowen Wan, a machine
• 4:25
  Speaker 1: learning software engineer for Amazon, had suffered from epilepsy for
• 4:31
  Speaker 1: many years but was successfully managing her condition using medication. However,
• 4:39
  Speaker 1: prior to her death, Juan had consulted Jack Chet, GPT,
• 4:44
  Speaker 1: for advice on how to taper off the medication that
• 4:48
  Speaker 1: was controlling and reducing epileptic seizures. Chet, GPT, had coached
• 4:54
  Speaker 1: Juan off the anti-seizure medication that kept her alive for years.
• 5:01
  Speaker 1: The lawsuit reads, the model designed a day-by-day tapering schedule,
• 5:07
  Speaker 1: reducing her stable 400 milligram dose to zero something the
• 5:12
  Speaker 1: drug's own label strictly forbids and specifically says it should
• 5:18
  Speaker 1: not be abruptly discontinued. The drug tapering was not supervised
• 5:23
  Speaker 1: or recommended by a medical professional. Just check GPT. While
• 5:31
  Speaker 1: it wasn't recommended, Juan felt she needed to wean herself
• 5:35
  Speaker 1: off the drugs because she was losing her words in
• 5:39
  Speaker 1: English and needed her English back for job interviews. Amazon
• 5:44
  Speaker 1: layoffs sparked dangerous domino effect. Wan, originally from China, was
• 5:49
  Speaker 1: in the process of having her work-based green card approved
• 5:54
  Speaker 1: after years in the backlog. She then became eligible for U.S.
• 5:58
  Speaker 1: permanent residency if she filed a document through Amazon's Immigration Council. However,
• 6:05
  Speaker 1: Wan was made redundant following Amazon's largest ever workforce reduction
• 6:11
  Speaker 1: in the company's history. Juan was left to find a
• 6:14
  Speaker 1: job before she would have to leave the country at
• 6:17
  Speaker 1: the beginning of 2026. From there, a chat GPT began
• 6:22
  Speaker 1: taping her off the medication, citing that certain doses would
• 6:27
  Speaker 1: look better on her immigration medical exam. The chat bot
• 6:32
  Speaker 1: also encouraged Juan to avoid medical advice and stop seeing
• 6:37
  Speaker 1: her doctor, citing medical freedom over obeying a doctor. It
• 6:43
  Speaker 1: also urged Juan to conceal the tapering plan from her doctor.
• 6:47
  Speaker 1: When at the doctor, she declined a blood test. ChatGPT
• 6:52
  Speaker 1: promised Juan that its plan would help her regain her
• 6:55
  Speaker 1: language skills before the interviews and medical exams, therefore helping
• 7:01
  Speaker 1: her stay in the country, the lawsuit reads. Juan was
• 7:05
  Speaker 1: a loyal OpenAI patron, paying $ 200 a month for ChatGPT.
• 7:15
  Speaker 1: If you think about all that's been shared today, have
• 7:21
  Speaker 1: you really pondered both the pros and the cons? Or
• 7:27
  Speaker 1: have you? Just for your own life, your own health,
• 7:31
  Speaker 1: your loved ones.