GDSC Hult Weekly Update: May 13th - May 19th Newsletter
Apple’s New ChatGPT Deal—Here’s What It Means For iPhone Security
Forbes (May 13, 2024)
OpenAI and Google have been negotiating with Apple to integrate new AI features into iOS 18, set to debut at the Worldwide Developers Conference next month. A recently published Bloomberg report shows that it’s almost certain the tech giant will be using OpenAI services to accomplish that.
Apple plans to align its AI services with its core value of privacy. The company has developed powerful chips, like the M4 in the latest iPads, to support on-device AI processing, keeping user data on the device rather than on external servers. This enhances data security and privacy. Apple is also integrating its Mac-grade chips into its servers to maintain control over data privacy for cloud-based AI capabilities.
Apple's "On-device LLM, Cloud-powered LLM, and Chatbot" strategy mirrors its long-standing partnership with Google for Search. While this approach ensures greater privacy, it may limit AI functionality on older Apple devices that lack the processing power needed for new AI capabilities. As noted by Forbes contributor Zak Doffman, only future applications and devices will fully benefit from iOS 18's AI features.
China's race to tech supremacy: Shenzhen speed
Financial Times (May 14, 2024)
In an episode of Financial Times’ podcast Techtonic, James Kynge, a veteran journalist on China, discusses how Shenzhen has become the engine of China's tech transformation. At the heart of the city, the Huaqiangbei district is known as China's Silicon Valley and houses the world's largest electronics market. Kynge interviews various people who highlight the market's extensive range of electronics, facilitating the creation of everything from everyday consumer electronics to advanced military systems.
Noah Zurking, an American developing an augmented reality headset in Huaqiangbei, emphasizes the advantage of immediate access to materials. This proximity enables China to compete with the US in technology and innovate independently, evolving quickly from imitator to innovator. The rapid pace of development in Shenzhen has propelled companies like DJI and BYD to the forefront of disruptive innovations.
According to the Australian Strategic Policy Institute, China leads in 37 of 44 critical technology areas, excelling in renewable energy, clean tech, electric vehicles, and quantum technologies. However, it still trails the US in generative AI, driven by major US tech companies. China also boasts the world's most popular app, TikTok, and leads in global patent filings, showcasing its rapid innovation.
Despite its authoritarian regime, China go against the conventional belief that innovation thrives best in democratic environments. The country's intense competitiveness, mentioned by the a Chinese venture capitalis Qi Zhou, and advanced manufacturing ecosystems, supported by skilled engineers, set it apart. Matt Sheehan, a fellow in the Asia Program of the Carnegie Endowment for International Peace, warns that the US cannot assume it will always lead in technology. If China maintains its current momentum, it could soon surpass the US, resulting in significant global impacts from an authoritarian tech leader.
Senate Group Recommends Spending Tens of Billions of Dollars on AI
Wall Street Journal (May 15, 2024)
The AI Working Group, led by Senate Majority Leader Chuck Schumer, released a report on Wednesday outlining a road map for AI policy. This report proposes tens of billions of dollars in new federal spending on AI safety and innovation to ensure the US remains competitive, particularly against China.
The proposal incorporates inputs from nine AI forums, addressing lawmakers' concerns about AI's impact on education, the military, and healthcare. Discussions involved around 150 AI experts, including tech CEOs, scholars, advocates, and labor representatives, on structuring national spending to foster AI development while mitigating its risks.
The report suggests a minimum increase of $32 billion in non-defense federal spending on AI. This funding would support initiatives such as the design and manufacturing of advanced AI chips, local election integrity efforts, and a series of “AI Grand Challenge” programs. Schumer emphasized the importance of this investment to stimulate innovation and deliver returns for taxpayers in the coming years, speaking to reporters on Tuesday evening.
As significant investments in AI are debated in places such as Silicon Valley and Washington, senators stress the need for a balance between fostering innovation and implementing regulations to prevent harmful societal disruptions. They specifically mention the regulation of opaque black box AI models as a critical area for legislative attention.
A ‘Digital Twin’ of Your Heart Lets Doctors Test Treatments Before Surgery
Wall Street Journal (May 16, 2024)
The concept of digital twins, long used in mechanical engineering, is now gaining traction in healthcare, with a market size valued at $1.6 billion in 2023. For instance, Johns Hopkins University professor Natalia Trayanova usescomputational heart models to predict areas prone to tiny scars, helping in the prevention of arrhythmia and future heart attacks. This is achieved through 3D modeling of enhanced MRI images, which guides the precise placement of new tiny scars using a procedure called ablation.
Digital twins, powered by AI, are expected to become widely accessible soon. Increased investment is driving their expansion across various sectors, potentially reducing harmful testing phases like drug trials. However, challenges include the complexity of human body modeling and privacy concerns with wearable device data.
Success stories include David Gakenheimer, an 80-year-old arrhythmia patient at Johns Hopkins, whose digital heart twin predicted real-life outcomes accurately. Other applications, such as modeling women's uteruses and cervixes for pregnancy predictions by Columbia University professor Kristin Myers, show digital twins' potential for improved diagnostics and testing. Experts at the National Cancer Institute see digital twins as a promising new approach to cancer treatment, potentially reducing harm and eliminating the uncertainties and additional treatments patients endure when going through “blind” drug testing.
Microsoft’s Partnership With Mistral AI Cleared by U.K. Regulator
Wall Street Journal (May 17, 2024)
After being under scrutiny for several weeks, Microsoft has finally received the green light from UK competition authorities regarding its investment in the French AI startup Mistral. The country’s Competition and Markets Authority (CMA) concluded that the deal does not warrant investigation under the Enterprise Act of 2002.
Not only Microsoft's investment but also Amazon's $4 billion stake have been under scrutiny for potential antitrust examination. Some companies have been finding loopholes to navigate regulations while gaining control over new technologies, as seen in "quasi-mergers," where they acquire control through investments or board seats without outright purchasing startups.
One possible reason why Microsoft passed scrutiny is that the nature of the partnership may not grant them significant influence, as noted by Alex Haffner, a competition partner at the UK law firm Fladgate.
In April, the CMA announced that it was soliciting third-party comments on partnerships between tech and artificial intelligence companies, including Microsoft and Amazon. As the CMA raised concerns about Big Tech's growing influence on the advanced AI market just last month, Amazon's fate still remains undisclosed pending a decision.
OpenAI created a team to control ‘superintelligent’ AI — then let it wither, source says
TechCrunch (May 18, 20224)
Led by former DeepMind researcher Jan Leike and OpenAI co-founder Ilya Sutskever, the Superalignment team was established last July with the aim of addressing the core technical challenges of controlling superintelligent AI within the next four years. It assessed the safety of both internal and external models and shared their findings. Initially, the team published extensive research and supported numerous researchers, but as the company's leadership shifted its focus more towards product launches, the team struggled to secure the necessary upfront investments.
Co-founder Sutskever played a key role in bridging the team with other sectors across the company and advocating for the importance of their work. However, he recently left the company following disagreements with CEO Sam Altman. Now, the team finds itself without the 20% of OpenAI computing resources promised to them, halting their progress and prompting several team members, including Leike, to resign this week. Leike cited changes in leadership priorities as the reason for his departure.
In response, Altman and co-founder Greg Brockman stated that there is still much more to be done in AI safety and that the company remains committed to this goal. John Schulman, another OpenAI co-founder, has taken on the responsibilities of the Superalignment team, but there will no longer be a dedicated team. It is crucial to closely monitor OpenAI's actions to ensure that they continue to prioritize both innovation and safety, and avoid moving in conflicting directions.
Thirst for AI knowledge drives growth in executive courses
Financial Times (May 20, 2024)
As AI tools like OpenAI’s ChatGPT gain traction, leaders are increasingly seeking to understand this transformative technology. A recent CarringtonCrisp survey of 10,000 learners across 40 countries found that about half plan to take AI-related courses within the next five years.
Business schools are meeting this demand by offering more AI, cybersecurity, and digital transformation courses. Executives like Annette Saller from Hartmann Group see the value of these programs, such as Insead’s Transforming Your Business with AI, which costs $2,008 and runs online for five weeks.
Prestigious institutions like NYU Stern, Iese Business School, London Business School, and MIT Sloan offer AI courses ranging from two-day seminars to multi-week programs, with prices from $3246 to $15,741. Even top executives need AI knowledge, but Julian Birkinshaw of London Business School warns that demand for new tech education can be cyclical, as seen with blockchain.
Developers must distinguish between hype-driven and essential AI course content, ensuring key aspects like IT infrastructure and data management are covered. The fast pace of AI advancements makes it hard to keep courses current and find qualified instructors. For example, MIT’s AI program involves professors from its Computer Science and Artificial Intelligence Laboratory.
Credits:
- Rafaella Andrade
- GANGA SINGH
- GDSC Events Team
References
- O'Flaherty, K. (2024, May 13). Apple's new ChatGPT deal: Here’s what it means for your iPhone. Forbes. https://www.forbes.com/sites/kateoflahertyuk/2024/05/13/apples-new-chatgpt-deal-heres-what-it-means-for-your-iphone/?sh=7a68d4cf583e
- Kynge, J. (2024, May 14.). Financial Times. Retrieved from https://www.ft.com/content/27b3c52a-a71d-424b-965a-772228c12b2f
- Hufford, A., & McKinnon, J. D. (2024, May 15). Senate group recommends spending tens of billions of dollars on AI. The Wall Street Journal. https://www.wsj.com/tech/ai/senate-group-recommends-spending-tens-of-billions-of-dollars-on-ai-4951279b?mod=ai_news_article_pos3
- Gormley, B., & Loftus, P. (2024, May 16). Digital heart surgery: A new era in patient treatment. The Wall Street Journal. https://www.wsj.com/health/healthcare/digital-heart-surgery-patient-treatment-c35ec4be?st=fpwg14qskfoxd7h
- Haffner, A. (2024, May 17). Microsoft's partnership with Mistral AI cleared by U.K. regulator. The Wall Street Journal. https://www.wsj.com/tech/microsofts-partnership-with-mistral-ai-cleared-by-u-k-regulator-cc473d6c?mod=tech_lead_pos2
- Lunden, I. (2024, May 18). OpenAI created a team to control superintelligent AI, then let it wither, source says. TechCrunch. https://techcrunch.com/2024/05/18/openai-created-a-team-to-control-superintelligent-ai-then-let-it-wither-source-says/
- Financial Times. (2024, May 19). Financial Times. Retrieved from https://www.ft.com/content/86b76229-bd48-4c07-8c8f-3b0cbf340a55