The rapid advancement of artificial intelligence is prompting significant shifts across various sectors. In education, Stanford professor Jure Leskovec has moved exams back to paper to better assess student understanding in the face of AI tools like GPT-3. Meanwhile, businesses are being urged to adopt generative AI to remain competitive, with adoption rates potentially mirroring the internet's growth, though leaders must also manage AI's impact on the workforce, including potential job displacement and the need for retraining. The National Institute of Standards and Technology (NIST) is developing new security guidelines, NIST SP 800-53 Control Overlays, to address the unique risks posed by AI systems. In the realm of AI development, OpenAI is reportedly partnering with Broadcom to create its own AI chips, aiming to reduce reliance on companies like NVIDIA, a move also being pursued by tech giants Google and Amazon. Oracle is leveraging AI to transform medicine, with its cloud infrastructure supporting large-scale AI projects and clients like OpenAI. The intersection of cryptocurrency and AI is also emerging, with blockchain technology and crypto payments creating marketplaces for specialized AI data, potentially accelerating development and creating new economic opportunities. However, legal challenges are mounting, as seen with Warner Bros. filing a copyright infringement lawsuit against AI image generator Midjourney for allegedly creating images of copyrighted characters. Concerns also persist about the underlying incentives in AI training that might contribute to AI 'hallucinations,' where models generate incorrect information.
Key Takeaways
- Stanford professor Jure Leskovec has switched to paper exams due to concerns about AI tools like GPT-3 affecting traditional assessment methods.
- Businesses are advised to adopt generative AI quickly to stay competitive, with AI adoption potentially following a rapid trajectory similar to the internet.
- Leaders face challenges managing AI's workforce impact, balancing efficiency gains with potential job displacement and the need for employee retraining.
- The National Institute of Standards and Technology (NIST) is proposing new security controls, NIST SP 800-53 Control Overlays, specifically for AI systems.
- OpenAI is reportedly collaborating with Broadcom to develop its own AI chips, aiming to decrease reliance on NVIDIA and other chip providers.
- Google and Amazon are also developing custom chips for AI, indicating a broader industry trend towards in-house hardware solutions.
- Oracle's AI technology is being used to advance medicine and enhance healthcare systems, supported by its robust cloud infrastructure.
- Cryptocurrency and blockchain technology are being explored as solutions for creating secure marketplaces for specialized AI data, addressing a key challenge in AI development.
- Warner Bros. has filed a copyright infringement lawsuit against AI image generator Midjourney, alleging the creation of images based on copyrighted characters.
- Questions are being raised about whether poor incentives in AI training might be contributing to AI 'hallucinations' or the generation of incorrect information.
Stanford professor moves exams to paper due to AI concerns
Stanford computer science professor Jure Leskovec has returned to paper exams for his large classes. This change came about two years ago as AI like GPT-3 became widely available. Students and teaching assistants felt traditional exams were no longer effective at testing knowledge. Leskovec agrees that grading these paper exams is more work but believes it's the best way to assess true understanding. He compares AI to a calculator, a powerful tool that requires learning how to use it and also testing independent thought.
Stanford professor switches to paper exams amid AI rise
Stanford computer science professor Jure Leskovec has reverted to paper exams for his classes due to the rise of AI tools. He explained that students and teaching assistants raised concerns about the effectiveness of current testing methods with AI's capabilities. While grading paper exams is more time-consuming, Leskovec believes it is the best approach to accurately gauge student knowledge. He likens AI to a calculator, emphasizing the need to test both the ability to use such tools and independent thinking.
NIST proposes security controls for AI systems
The National Institute of Standards and Technology (NIST) is developing new security guidelines specifically for artificial intelligence systems. These guidelines, called NIST SP 800-53 Control Overlays, aim to address the unique risks AI presents. They will cover various AI applications and ensure security is part of the development process. NIST has also launched a Slack channel for experts to collaborate and provide feedback on these evolving security measures.
Crypto fuels AI data needs, creating new marketplaces
Cryptocurrency and AI are merging, with crypto poised to solve AI's biggest challenge: data. While computing power and algorithms are dominated by large companies, data is a decentralized resource. Crypto payments and blockchain technology can create secure and efficient marketplaces for specialized AI data. This allows individuals worldwide to contribute data and be rewarded, potentially accelerating AI development and creating new economic opportunities.
OpenAI plans in-house AI chip with Broadcom
OpenAI is reportedly planning to develop its own AI chips in partnership with Broadcom, aiming for a release next year. This move would reduce OpenAI's dependence on companies like NVIDIA for essential computing power. Tech giants such as Google and Amazon are also creating custom chips for AI. Broadcom has announced significant AI infrastructure orders from a new customer, fueling speculation that OpenAI is behind it. This initiative is a key step for OpenAI to control its own hardware infrastructure.
AI offers businesses a wake-up call for adoption
Generative AI is rapidly evolving, and businesses need to adopt it to stay competitive. The author shares personal experiences using AI tools like ChatGPT for content creation and image generation. While AI can be a powerful time-saver, it's important to use it effectively and be aware of its limitations. The article compares AI adoption to the internet's growth, suggesting a faster pace of change. Businesses that don't integrate AI risk falling behind, and leaders must consider its impact on the workforce and human intellect.
Warner Bros. sues AI image generator Midjourney
Warner Bros. Discovery, along with DC Comics and other subsidiaries, has filed a copyright infringement lawsuit against the AI image generator Midjourney. The lawsuit alleges that Midjourney allows users to create images of copyrighted characters like Batman and Superman, infringing on their intellectual property. Warner Bros. claims Midjourney knowingly profits from this piracy. This follows similar lawsuits filed by Disney and Universal against the AI firm, highlighting ongoing legal battles over AI's use of copyrighted material.
Are poor incentives causing AI hallucinations?
This article explores whether flawed incentives might be contributing to AI 'hallucinations,' where AI systems generate incorrect or nonsensical information. It suggests that the way AI models are trained and rewarded could inadvertently encourage such errors. The piece prompts a deeper look into the underlying motivations and reward structures within AI development to address this growing issue.
Oracle AI transforming medicine and saving lives
Oracle is developing artificial intelligence that is revolutionizing medicine and helping to save lives, according to Yael Har Even, SVP and Country Manager of Oracle Israel. Oracle's AI technology is enhancing the health system, enabling doctors to communicate with patients more effectively. Israel serves as a key innovation hub for Oracle in cloud computing and AI. The company's efficient cloud infrastructure is crucial for AI models, attracting major clients like OpenAI and supporting large-scale projects such as the U.S. government's Stargate initiative.
Leaders face challenges managing AI's workforce impact
AI is set to significantly change the nature of work, and leaders face challenges in managing its impact on the workforce. While AI may displace some jobs, it also creates new opportunities in fields like data engineering. Companies are adopting AI for efficiency, but this can lead to job losses and affect company culture. Boards of directors must balance AI adoption with their responsibility to employees, considering support for retraining and transparent communication about AI plans.
Sources
- This Stanford computer science professor went to written exams 2 years ago because of AI. He says his students insisted on it
- This Stanford computer science professor went to written exams 2 years ago because of AI. He says his students insisted on it
- NIST Releases Control Overlays For Securing AI Systems
- Crypto Is Powering The AI Age’s Most Critical Fuel: Data
- OpenAI is preparing to enter the semiconductor race, with plans to roll out its first in-house AI chip next year in partnership with US chip giant Broadcom, according to a Financial Times report. The…
- Jodi's Journal: The case for an AI wake-up call
- Warner Bros., DC Comics and More Sue Midjourney, Says AI Firm 'Thinks It Is Above the Law'
- Are bad incentives to blame for AI hallucinations?
- "Oracle is developing artificial intelligence that is reshaping medicine and helping
- Leadership Challenges In Addressing AI’s Inevitable Workforce Impact