| |
| 1 | | OpenAI CEO Sam Altman fielded questions from Congress today, where he expressed support for licensing and safety standards for more advanced AI systems. During the congressional hearing, the CEO of ChatGPT's AI company emphasized the importance of government intervention to address the risks tied to increasingly powerful AI systems. The session marked the first in a series of anticipated hearings on AI as lawmakers from across the political spectrum consider future AI regulations. More: - During Tuesday's Senate Judiciary subcommittee hearing, Altman voiced concerns about the potential for substantial harm caused by AI.
- "I think if this technology goes wrong, it can go quite wrong, and we want to be vocal about that," he said, adding that the industry wants "to work with the government to prevent that from happening."
- He also acknowledged that GPT-4 would automate some jobs entirely but also create some new jobs.
- "GPT4 and other systems like it are good at doing tasks, not jobs, so you see already people that are using GPT4 to do their job much more efficiently."
- Altman proposed a three-point plan, including the creation of safety standards to evaluate higher-tier AI models and a government agency to license those models — and take away licenses if not in compliance.
Risks of AI: - To highlight the risks of unregulated AI, Sen. Richard Blumenthal (D-Conn.) opened the hearing by playing a fake recording of his voice, written by ChatGPT and cloned with AI using audio from his floor speeches.
- Blumenthal expressed fears that ChatGPT could be used to generate content to endorse Ukraine's surrender or express support for Vladimir Putin's leadership, for example.
- He and other lawmakers acknowledged that Congress failed to regulate social media early enough, resulting in toxic content and harm to younger people and children, among other problems.
Further concerns: - Lawmakers from both political parties questioned Altman, along with the other witnesses — Gary Marcus, an emeritus professor at NYU, and Christina Montgomery, IBM's VP and chief privacy and trust officer — about various AI risks.
- Their concerns about generative AI included job disruption, election misinformation, copyright and liability issues, harmful content, and impersonation.
- IBM's Montgomery stressed the importance of regulating these risks rather than the technology itself, cautioning against a reckless approach.
- Marcus said hyper-targeted advertising would become common in the AI industry, despite assurances from OpenAI and IBM. He called for a dedicated Cabinet-level organization in the U.S. to oversee AI due to the risks and information complexity.
|     | |
| 2 | | What happened: Hippocratic AI, a healthcare-focused generative AI startup, launched today with $50M in seed funding. The company is building large language models specifically for healthcare tasks, such as offering dietary advice, answering pre-op and billing questions, and delivering negative test results. Why it matters: Unlike other AI models in healthcare, Hippocratic AI's focus is not on diagnosing medical conditions. Its consumer-facing models could power healthcare chatbots that supplement the roles of clinicians, billing specialists, and others as the healthcare industry copes with burnout and high turnover. What the numbers say: Hippocratic AI tested its AI model on 114 healthcare certifications, outperforming OpenAI's GPT-4 on 105 of them, according to co-founder and CEO Munjal Shah. These certifications spanned various specialties, including urology, pharmacy, and nursing. What's next: Andreessen Horowitz, one of the investors, said Hippocratic was one of the largest seed checks it has made in the healthcare industry. The funding will be used to support model training and safety as the startup seeks to differentiate itself in the healthcare sector, integrating medical expertise and incorporating feedback from doctors in its model development. The bigger picture: Hippocratic's launch came on the same day that the World Health Organization (WHO) cautioned against the use of AI in healthcare due to potential bias and misinformation. While recognizing the benefits of AI in improving access to health information, WHO emphasized the need to evaluate the risks associated with using large language model tools like ChatGPT to protect human well-being and public health. 
| | | |
| A message from our sponsor, TRUERA. |  | | ML Monitoring at Scale – the MLOps Perspective Ready to see what it really takes to monitor machine learning models in production? This in-depth webinar takes you up close and personal. Join TruEra on May 23rd as their Chief Scientist, Anupam Datta, and Biswajit Das, the former Chief Architect of the AI Platform at a large financial services firm, provide a real-life overview of the fundamentals of ML Monitoring. During the 60-minute webinar, they will cover: - What is ML Monitoring/ML Observability?
- Monitoring best practices - going beyond uptime and retraining
- Key elements of an ML monitoring platform: scale, integrations, and enterprise readiness
- Case study: monitoring and MLOps in real life
- Live, practical run-through of how to identify, diagnose, and debug model issues in production
- Q&A with the presenters
Date: May 23rd Where: Virtual event - Join from Anywhere
Who should attend: Those who manage ML operations, data engineering, data science, or AI Join for free | |
|
| 3 | | Microsoft researchers recently published a research paper on GPT-4, claiming that it exhibits "sparks" of artificial general intelligence (AGI) or cognitive abilities on par with humans. GPT-4 performs tasks remarkably close to human-level performance in various domains such as mathematics, coding, vision, medicine, law, and psychology, according to the researchers. More: - The paper, titled "Sparks of Artificial General Intelligence: Early experiments with GPT-4," highlights GPT-4's potential as an early version of an AGI system while acknowledging the need to identify its limitations and overcome challenges in advancing towards more comprehensive AGI systems.
- The researchers were particularly impressed with GPT-4's solution to a reasoning exercise involving stacking objects.
- They asked both AI models to stack a book, nine eggs, a laptop, a bottle, and a nail in a stable way. While GPT-3 seemed confused, GPT-4 surprised the researchers with its ingenious approach.
- GPT-4 suggested arranging the eggs in a three-by-three grid on top of the book, creating a stable platform for the laptop and other objects to balance.
- It explained that the laptop would securely fit within the book and eggs, providing a stable surface for the next layer.
- "All of the things I thought it wouldn't be able to do? It was certainly able to do many of them — if not most of them," lead author Sébastien Bubeck told The New York Times.
| | | |
| 4 | | Apple plans to introduce an AI-powered voice cloning feature for iPhones, iPads, and Macs. The upcoming "Personal Voice" tool generates synthetic versions of a user's voice for phone and FaceTime calls. It works with Apple's new AI Live Speech text-to-speech feature, allowing users to type and have the words spoken out loud. More: - The tool, which replaces Siri with an automated version of the user's voice, is aimed at people who are at risk of losing their speaking abilities.
- To create a Personal Voice, users read a series of text prompts aloud into their microphone, totaling 15 minutes of audio recorded on their device.
- The audio is processed locally on Apple devices (iPhones, iPads, or Macs) using machine learning.
- The result is a generated voice that sounds like the sender's voice
Zoom out: - Apple today revealed other accessibility enhancements for people with cognitive, vision, hearing, and mobility impairments, including a mode that combines camera and LiDAR input with machine learning to announce on-screen text.
- The features are expected to roll out with iOS 17, the next major iPhone update.
| | | |
| A message from our sponsor, WING VC. |  | | Meet the Best Companies in Enterprise Tech: Presenting the 5th Annual ET30 The Enterprise Tech 30, now in its 5th year, identifies the best companies in enterprise tech as chosen by an exclusive group of 96 venture capitalists across 74 firms. It serves as a platform for the tech industry to identify, learn about, and build relationships with today’s most promising private companies. This year’s cohort is unlike any of the previous four. In this report you can expect: - Over two-thirds are new to the ET30.
- Nearly 30 percent of the companies on the list are in the generative AI category.
- Half are in the developer platform and tools category, an increase from years past.
- Many span multiple categories, including security, fintech, and productivity tools.
And, some of the previous year’s winners are seeing tremendous success: Of last year’s ET30 companies, one has been acquired (pending), and 20 have raised follow-on capital of $9.3 billion in total. Of the initial ET30 companies in 2019, the total valuation has increased by $264 billion since the 2019 report publication date. See The List | |
|
| |
| |
| Upcoming Events | | MAY
17 | | Monthly Meditation guided by Nicholas Whitaker | | | | | | | MAY
23 | | In-depth webinar on ML Monitoring at Scale with TruEra | | | | | | | MAY
23 | | Advance your security program and win more deals by joining this Vanta webinar | | | | | | | MAY
25 | | Inside.com Book Club - Zero To One by Peter Thiel | | | | | | * This is a sponsored event | | | |
Term of the Day Early adopters: Early adopters are the first users of a startup’s products and services. They are key in giving feedback on the business proposition of entrepreneurs. Question of the Week How do you approach networking, and what advice can you give others looking to build their professional connections? Join the conversation |
INSIDE AI LEADERBOARD (7 DAYS) |
 | AI and technology writer | | Beth is a writer and editor at Inside.com covering artificial intelligence and daily technology news. She has written the Inside AI newsletter since 2019. You can reach her at Beth.Duckett@inside.com. | | This newsletter was edited by Aaron Crutchfield | |
| |
| |
|
 | The definitive list of the best in enterprise tech, by company size, download now. | |
|
| |
Comments
Post a Comment