cover of episode #116 - ChatGPT plugins, AI hardware, petition to pause AI, Trump deepfakes

#116 - ChatGPT plugins, AI hardware, petition to pause AI, Trump deepfakes

2023/3/31
logo of podcast Last Week in AI

Last Week in AI
AI Deep Dive AI Chapters Transcript
People
A
Andrey Kurenkov
J
Jeremy Harris
Topics
Andrey Kurenkov：OpenAI为ChatGPT添加插件支持，使其能够访问互联网和外部服务，极大地扩展了其功能。这带来了新的机遇和挑战，插件的安全性需要谨慎考虑。 Andrey Kurenkov：NVIDIA的Hopper GPU（H100）针对Transformer模型进行了优化，显著提高了AI训练效率，这标志着NVIDIA在AI硬件领域的领先地位。 Andrey Kurenkov：高盛的研究表明，生成式AI将对3亿个工作岗位产生影响，其中7%的工作岗位将至少有一半的工作内容被AI完成。这将对全球经济产生深远的影响。 Andrey Kurenkov：Agility Robotics的Digit机器人被设计用于仓库中搬运塑料箱，标志着人形机器人向实际应用迈进。 Andrey Kurenkov：Runway Gen-2是第一个公开可用的文本转视频生成器，标志着AI在视频生成领域的进展。 Andrey Kurenkov：Clearview AI的面部识别技术已被美国警方使用近100万次，这引发了人们对隐私和安全问题的担忧。 Andrey Kurenkov：澳大利亚Centrelink使用的语音识别系统容易受到AI的欺骗，这凸显了AI技术带来的安全风险。 Andrey Kurenkov：AI技术进步使得生成逼真的图像变得更容易，这增加了深度伪造的风险。 Andrey Kurenkov：AI生成的特朗普被捕照片并未广泛欺骗公众，这表明公众对AI生成的图像仍保持警惕。 Andrey Kurenkov：美国编剧协会（WGA）允许在编剧中使用人工智能，但要求维护编剧的署名权。 Jeremy Harris：ChatGPT插件带来的新工具，会产生新的能力和风险，这是一个值得关注的AI系统演变阶段。为AI系统添加新工具会带来新的机遇和风险，这与单纯扩大模型规模观察其新能力不同，是一个值得研究的‘阶跃式’变化。 Jeremy Harris：NVIDIA推出Hopper GPU（H100），针对Transformer模型进行了优化，并提供云服务，进一步推动了AI发展。 Jeremy Harris：OpenAI的Ilya Sutskever认为，目前AI发展主要集中在工程问题上，而非算法突破。 Jeremy Harris：超过1100名知名人士签署了一封公开信，呼吁所有AI实验室暂停至少6个月的时间，以评估AI技术的风险。 Jeremy Harris：旧金山的一个博物馆关注AI的潜在风险，旨在提高公众对AI安全性的认识。 Jeremy Harris：AI技术进步使得生成逼真的图像变得更容易，这增加了深度伪造的风险。

Deep Dive

Chapters
Discussion on how OpenAI's addition of plugin support to ChatGPT will allow it to interact with external services, potentially leading to new capabilities and risks.

Shownotes Transcript

Our 116th episode with a summary and discussion of last week's big AI news!

Check out our text newsletter at https://lastweekin.ai/)

Stories this week:

Applications & Business OpenAI is massively expanding ChatGPT’s capabilities to let it browse the web and more)

NVIDIA's big AI moment is here)

Lighting Round Former head of Google China joins ChatGPT frenzy by starting own venture)

Cerebras Systems Releases Seven New GPT Models Trained on CS-2 Wafer-Scale Systems)

Generative AI set to affect 300 million jobs across major economies)

Agility’s Latest Digit Robot Prepares for its First Job)

Research & Advancements Learning to grow machine-learning models)

Researchers from UC Berkeley and Deepmind Propose SuccessVQA: A Reformulation of Success Detection that is Amenable to Pre-trained VLMs such as Flamingo)

Lighting Round New virtual testing environment breaks the 'curse of rarity' for autonomous vehicle emergency decision-making) 

Scientists are using machine learning to forecast bird migration and identify birds in flight by their calls )

New in-home AI tool monitors the health of elderly residents)

Runway Gen-2 is the First Publicly Available Text-to-Video Generator)

Policy & Societal Impacts

OpenAI co-founder on company’s past approach to openly sharing research: ‘We were wrong’)

1,100+ notable signatories just signed an open letter asking “all AI labs to immediately pause for at least 6 months”)

Lighting Round In San Francisco, some people wonder when A.I. will kill us all)

Clearview AI used nearly 1m times by US police, it tells the BBC

Voice system used to verify identity by Centrelink can be fooled by AI)

AI can draw hands now. That’s bad news for deep-fakes)

Art & Fun Stuff

People Aren’t Falling for AI Trump Photos (Yet))

WGA Would Allow Artificial Intelligence in Scriptwriting, as Long as Writers Maintain Credit)