Latest / Elon Musk Podcast / Why would anyone upload medical data to Elon's X?

Transcript

Why would anyone upload medical data to Elon's X? · Elon Musk Podcast

• 0:01
  Hey everybody. Welcome back to the Elon Musk
• 0:04
  Podcast. This is a show where we discuss
• 0:07
  the critical crossroads that shape SpaceX, Tesla X, The
• 0:11
  Boring Company, and Neurolink. I'm your host, Will Walden.
• 0:16
  So what happens when a tech giant like Elon Musk encourages
• 0:19
  users to share their medical images with an artificial
• 0:23
  intelligence tool? Now, could Grok revolutionize
• 0:26
  diagnostics, or does it pose a serious privacy risk?
• 0:31
  And why are privacy experts raising alarms about this new
• 0:35
  experiment on X? So over the past weeks, users of
• 0:39
  the X platform have been uploading personal medical
• 0:43
  scans, X-rays, Mris, and more to Grok, an AI chat bot developed
• 0:49
  by Musk's company X AI. They're asking for analysis,
• 0:54
  seemingly eager to test Musk's claim that Grok is poised to
• 0:58
  become extremely good in interpreting medical data.
• 1:02
  But this move has sparked sharp debates among privacy advocates,
• 1:05
  medical professionals, and AI experts.
• 1:08
  Now Elon Musk introduced Grok as a cutting edge AI system
• 1:11
  designed to offer fast, witty and sometimes provocative
• 1:15
  answers leveraging real time knowledge from the X platform.
• 1:19
  It was first unveiled in late 2023 and by mid 2024, Grok had
• 1:24
  advanced to handling complex visual inputs, including medical
• 1:28
  images. And Musk's Reek sent
• 1:30
  encouragement to submit health scans has added a new dimension
• 1:34
  to its functionality, but also intensified scrutiny.
• 1:38
  Now, unlike more conventional medical AI systems developed in
• 1:42
  collaboration with hospitals or researchers, Grok was trained on
• 1:46
  public social media data, a method that raises ethical
• 1:49
  concerns. And Musk pitched the chatbot as
• 1:52
  a tool for users to seek medical insights, either as a
• 1:56
  preliminary analysis or a second opinion.
• 2:00
  Now, Musk said, this is still early stage but is already quite
• 2:03
  accurate and will become extremely good.
• 2:06
  This openness has attracted both laypeople and professionals
• 2:10
  curious to see if the AI can match human expertise.
• 2:14
  Now, reports from users show a range of outcomes.
• 2:16
  Some praise Grok for his ability to recognize medical conditions.
• 2:20
  One individual said, had it checked out my brain tumor, not
• 2:23
  bad at all. While others highlighted glaring
• 2:25
  errors, like mistaking a broken clavicle for a dislocated
• 2:29
  shoulder. Physicians have also tested the
• 2:32
  tool. An immunologist from the Jackson
• 2:35
  Laboratory noted success after tweaking his input prompt.
• 2:39
  The response is very thorough and impressive, though, he said,
• 2:42
  describing Grok's diagnosis of a X-ray now.
• 2:46
  Others, like Doctor Laura Heacock, a radiologist and
• 2:50
  researcher at NYU Imaging, were less satisfied.
• 2:53
  Testing Grok with a set of breast radiology images, Heacock
• 2:57
  reported 0 accurate diagnosis. In one case, the AI mistook an
• 3:02
  obvious cancerous mass for calcifications.
• 3:06
  For now, non generative AI methods continue to outperform
• 3:09
  in medical imaging, Heacock said.
• 3:12
  Now, privacy concerns are a big deal.
• 3:15
  Musk's invitation to share sensitive medical data has
• 3:18
  sparked outrage among privacy experts.
• 3:21
  Medical information uploaded to Grok falls outside the
• 3:24
  protection of the Health Insurance Portability and
• 3:27
  Accountability Act HIPAA, and HIPAA governs data handled by
• 3:33
  healthcare professionals, providers, insurers and their
• 3:37
  partners, but it doesn't extend to social media platforms like X
• 3:40
  or their affiliated AI tools. Brandy Malin, a biomechanical
• 3:47
  Sorry, a biomedical informatics professor at Vanderbilt
• 3:51
  University, criticized the casual nature of data sharing
• 3:54
  with Grok. He said posting personal
• 3:57
  information to Grok is more like, we let's throw this data
• 4:00
  out there and hope the company is going to do what I want them
• 4:03
  to do, he remarked. XS privacy policy states that
• 4:06
  while the company doesn't sell user data to third parties, it
• 4:10
  does share information with related entities.
• 4:13
  This ambiguity fuels fears about how the data might be used.
• 4:16
  Matthew McCoy, an assistant professor of medical ethics at
• 4:20
  the University of Pennsylvania, expressed skepticism about the
• 4:23
  tool safeguards, saying, As an individual user, would I feel
• 4:27
  comfortable contributing health data?
• 4:30
  Absolutely not. You know, sharing health data on
• 4:32
  platforms like X risks creating a permanent digital footprint
• 4:37
  that could lead to unintended consequences.
• 4:40
  Consider a PET scan showing early signs of Alzheimer's
• 4:44
  disease. If linked to a user's online
• 4:46
  identity, such data could be exploited by future employers,
• 4:50
  insurers, or even housing associations.
• 4:53
  While laws like the Genetic Non Discrimination Act offers some
• 4:57
  protection, loopholes remain for entities like life insurance
• 5:01
  providers. Moreover, there's the question
• 5:03
  of how Grok itself uses the data.
• 5:05
  The company claims that users control whether their inputs are
• 5:08
  used to improve the AI, but experts warn that most users
• 5:12
  lack the technical literacy to fully understand or manage these
• 5:16
  settings. Now, even with safeguards in
• 5:18
  place, the risk of data breaches or misuse loom large.
• 5:22
  Beyond privacy, Grok's medical accuracy remains a pressing
• 5:26
  issue. Experts stress that training a
• 5:29
  robust AI for healthcare requires high quality, diverse
• 5:33
  data sets and deep collaboration with medical professionals.
• 5:36
  Without these elements, AI systems risk producing
• 5:39
  unreliable or even dangerous outputs.
• 5:42
  SU Chi Surya, director of the Medicine, Learning and
• 5:46
  Healthcare Lab at John Hopkins Johns Hopkins University,
• 5:51
  compared untrained AI tools to a hobbyist chemist mixing
• 5:55
  ingredients in the kitchen sink. She said that the potential for
• 5:58
  harm if users act on inaccurate diagnosis leads to unnecessary
• 6:03
  tests or treatments. Dr. Soraya also highlighted the
• 6:07
  significant challenges of integrating AI like Grok into
• 6:11
  clinical systems, where rigorous validation and adherence to
• 6:15
  ethical guidelines are paramount now.
• 6:18
  While tools like Grok might enhance workflows in the future,
• 6:21
  the current state lacks the recision required for critical
• 6:25
  healthcare applications. Grok has already attracted
• 6:28
  regulatory attention in Europe. Privacy regulators have flagged
• 6:32
  the tool for potential violations of the General Data
• 6:35
  Protection Regulation, GDPR. The Irish Data Protection
• 6:39
  Commission recently petitioned the courts to stop X from using
• 6:42
  social media data to train Grok, citing concerns over data
• 6:46
  harvesting practices in the US. Grok faces scrutiny from
• 6:51
  government officials worried about a is role in spreading
• 6:54
  misinformation. Minnesota Secretary of State
• 6:57
  even urged X to direct election related queries to verified
• 7:02
  sources like Can I vote.org Now? Musk's response to these
• 7:06
  criticisms has been dismissive. In August, XAI called European
• 7:10
  regulators actions unwarranted and defended its decision to let
• 7:13
  users opt in or out of data sharing.
• 7:16
  Unlike the rest of the AI industry, we choose to provide a
• 7:19
  simple control to all X users, the company stated, positioning
• 7:23
  itself as a defender of user autonomy.
• 7:26
  Now, Musk's interest in AI extends beyond Grok, though.
• 7:29
  His other ventures, including Neurolink, have a broader
• 7:33
  ambition to merge human intelligence and technology.
• 7:36
  Neurolink recently received FDA approval for a brain computer
• 7:40
  interface aimed at restoring vision in blind individuals.
• 7:44
  In such developments demonstrate Musk's commitment to positioning
• 7:48
  AI at the forefront of healthcare innovation.
• 7:51
  And despite these advancements, though Musk's approach has often
• 7:54
  clashed with ethical concerns, his long standing warnings about
• 7:58
  the existential risks of AI contrast sharply with his
• 8:02
  promotion of tools like Rock, which critics argue are being
• 8:06
  deployed without significant oversight now.
• 8:10
  For now, Grok offers intriguing possibilities but also
• 8:14
  significant risks. Users should exercise caution
• 8:18
  before uploading sensitive medical data, especially given
• 8:21
  the uncertainties around privacy and accuracy.
• 8:24
  And as the tool evolves, though, it may become a more reliable
• 8:27
  resource. But for now, it remains an
• 8:29
  experimental platform with serious ethical and practical
• 8:34
  limitations. Hey, thank you so much for
• 8:41
  listening today. I really do appreciate your
• 8:43
  support. If you could take a second and
• 8:45
  hit this subscribe or the follow button on whatever podcast
• 8:48
  platform that you're listening on right now, I greatly
• 8:51
  appreciate it. It helps out the show
• 8:53
  tremendously and you'll never miss an episode.
• 8:55
  And each episode is about 10 minutes or less to get you
• 8:59
  caught up quickly. And please, if you want to
• 9:01
  support the show even more, go to patreon.com/stagezero and
• 9:08
  please take care of yourselves and each other and I'll see you
• 9:10
  tomorrow.