News from OpenAI and the world isn't ending?
Hello there, AI enthusiasts! It's been another bustling week in the world of artificial intelligence, and we're here to bring you the latest and greatest updates surrounding the two giants in the field: OpenAI and Google. Let's dive into the stories that are shaping the future of AI and end on a serious question which we would love your input on.
Elon Musk's xAI nets $6 Billion to Build Skynet's Friendlier Cousin, Promises Not to End Humanity
Published by VentureBeat, Elon Musk's xAI startup has secured a massive $6 billion in funding, affirming its aggressive stance in the AI industry with a pre-money valuation of $18 billion.
The funding will fuel xAI's competition with major players such as OpenAI and Google, accelerating the development of its Grok-1 model and chatbot towards the goal of artificial generalized intelligence (AGI).
xAI, having already made significant strides with Grok-1 and its successors Grok-1.5 and Grok-1.5V, plans to invest in GPU clusters and talent to enhance its AI systems, positioning itself as a formidable rival to OpenAI's ChatGPT and Google's Gemini.
OpenAI's New Safety and Security Committee
OpenAI has established a new Safety and Security Committee (SSC) to guide critical decisions as it advances towards the creation of the successor to GPT-4, an initiative reported by The Register.
The SSC, including CEO Sam Altman and other board members, will play a pivotal role in ensuring the safe development of OpenAI's "next frontier model," amid the organization's recent high-profile personnel changes.
The formation of this safety-focused group follows the departure of key figures such as co-founder Ilya Sutskever and the disbanding of the Superalignment group, signaling OpenAI's commitment to addressing AI safety and security concerns.
Phew - we can all stop worrying about it then. :-)
GPT-4: A New Era in Financial Analysis
Researchers at the University of Chicago have discovered that large language models like GPT-4 can analyze financial statements with a precision that rivals or surpasses human analysts, as reported by VentureBeat.
GPT-4, utilizing innovative "chain-of-thought" prompts, achieved a remarkable 60% accuracy in predicting earnings growth, surpassing the 53-57% accuracy range of professional analysts, heralding a new era in financial analysis.
The study's findings suggest that LLMs, with their vast knowledge bases and pattern recognition capabilities, may soon play a pivotal role in financial decision-making, as detailed in the paper from the University of Chicago.
OpenAI and Publishers: A Strategic Partnership
OpenAI has forged a "multi-year" and "historic" partnership with News Corp, allowing it to train its AI models on content from esteemed publications such as The Wall Street Journal and Financial Times, as reported by TechCrunch.
The collaboration with News Corp also grants OpenAI rights to feature News Corp mastheads in ChatGPT, marking a significant PR victory for the AI company and financial benefits for the publisher.
Although these content licensing deals with OpenAI have an expiration date, they offer a strategic pause for both parties, allowing future renegotiations in an industry where perpetual licensing is uncommon.
GDPR Compliance: The EU Scrutinizes OpenAI's ChatGPT
The EU's data protection taskforce has been analyzing how GDPR applies to OpenAI's ChatGPT, with initial findings showing uncertainty around the legality of data processing, as reported by TechCrunch.
Amidst GDPR compliance investigations, OpenAI faced a temporary ban in Italy, which lifted only after the company addressed the Italian DPA's concerns, yet the legality of data processing for AI training remains under EU scrutiny.
While OpenAI could face penalties up to 4% of global annual turnover for GDPR violations, the report suggests that with adequate safeguards and transparency, the company might adjust its data processing practices to comply with EU regulations.
OpenAI CEO Assures Staff Their Stock Options are Safe, Still No Word on Office Snack Situation
OpenAI's CEO, Sam Altman, assured employees that vested equity would remain untouched, responding to concerns raised by a Vox report over restrictive off-boarding agreements, as covered by The Register.
In a positive turn of events, OpenAI management informed both former and current staff that no vested equity had been or would be canceled, rectifying the issue highlighted in the leaked documents seen by Vox.
But never fear VOX and OpenAI look to be on good terms again :-).
Add this to the list of Reddit, The Atlantic, News Corp, Stack Overflow and the Financial Times.
AI Trained on Tabloid News Shocked to Learn World Isn't Actually Ending Every Week
When AI models are trained on news from companies which mixes news with entertainment, we need to think about what that means for the AI's understanding of the world. These companies often prioritize catchy, engaging content, which might not always be the most accurate or serious.
So, if AI starts learning from these sources, there's a risk it could adopt some of those biases or inaccuracies. It's like teaching someone using only blockbuster movies as textbooks; they might end up with a very dramatic, but not entirely accurate, view of the world.
As we develop more advanced AI, we should think carefully about where the information comes from and whether it represents the values and truths we want to uphold.
