Last Week in AI #118
Disinformation, the OpenAI Startup Fund, and more!
One major concern about GPT-3 and mature generative AI systems is their ability to create images, text, and other forms of information that may conceivably been created by humans. Generated media could be used to mislead people, impact national politics, and have a number of other negative consequences. A group of researchers at Georgetown found that these concerns aren't just theoretical: the researchers used GPT-3 to generate various sorts of misinformation and found that, with some human curation, GPT-3 was quite effective at promoting falsehoods.
On May 26, OpenAI launched the OpenAI Startup Fund. Managed by OpenAI with investment from Microsoft and other partners, the $100 million fund aims to "help AI companies have a profound, positive impact on the world." According to OpenAI CEO Sam Altman, the fund will make "big, early bets" on very few companies. In particular, it looks to partner with early-stage startups in areas where AI can have a "transformative" effect. Such fields include healthcare, climate change, and personal assistance.
Advances & Business
Chinese startup Pony.ai gets approval to test driverless vehicles in California - "Chinese autonomous vehicle startup Pony.ai has received a permit from California’s Department of Motor Vehicles to test its driverless cars without human safety drivers behind the wheel on specified streets in three cities."
An AI has disproved five mathematical conjectures with no human help - "An artificial intelligence has disproved five mathematical conjectures, despite not being equipped with any information about the problems."
The Costly Pursuit of Self-Driving Cars Continues On. And On. And On. - "Many in Silicon Valley promised that self-driving cars would be a common sight by 2021. Now the industry is resetting expectations and settling in for years of more work."
DeepMind Presents Neural Algorithmic Reasoning: The Art of Fusing Neural Networks With Algorithmic Computation - "Algorithms are everywhere. They detail the specific instructions that computers need to carry out tasks, from self-driving vehicles to recommendation systems and even in your microwave oven."
Facebook’s Dynabench now scores NLP models for metrics like ‘fairness’ - "Elevate your enterprise data technology and strategy at Transform 2021. Last September, Facebook introduced Dynabench, a platform for AI data collection and benchmarking that uses humans and models “in the loop” to create challenging test datasets."
Yoshua Bengio Team’s Recurrent Independent Mechanisms Endow RL Agents With Out-of-Distribution Adaptation and Generalization Abilities - "One of the interesting and enduring challenges in machine learning (ML) is improving model capabilities for out-of-distribution adaptation and generalization."
Tesla announces transition to 'Tesla Vision' without radar, warns of limitations at first - "Tesla today announced the official transition to “Tesla Vision” without radar on Model 3 and Model Y. In the process, the automaker warns of some limitations on Autopilot features at first."
Terrible news, everyone: AI is learning how to post cringe - "People like to argue that technology is value neutral; that it’s neither inherently good nor inherently bad, but can simply be put to different uses."
AI Could Soon Write Code Based on Ordinary Language - "In recent years, researchers have used artificial intelligence to improve translation between programming languages or automatically fix problems. The AI system DrRepair, for example, has been shown to solve most issues that spawn error messages."
Microsoft's first OpenAI-powered feature helps beginners build productivity apps - "Microsoft has officially introduced its first GPT-3-powered feature in a customer product, eight months after it exclusively licensed the sophisticated OpenAI language model."
Sony’s AI subsidiary is developing smarter opponents and teammates for PlayStation games - "In 2019, Sony quietly established a subsidiary dedicated to researching artificial intelligence. What exactly the company plans to do with this tech has always been a bit unclear, but a recent corporate strategy meeting offers a little more information."
Baidu debuts updated AI framework and R&D initiative - "Elevate your enterprise data technology and strategy at Transform 2021. At its bi-annual deep learning conference, Wave Summit, Baidu announced version 2.1 of PaddlePaddle, the company’s framework for AI and machine learning model development."
Computers Will Be Able to Read Images From Your Brain Within a Decade - "I have a photographic memory, and I’m a time-space synesthete. That means I can visualize basically any place I’ve ever been, in photorealistic detail. I can also imagine non-existent places, and fly around them in my brain like I’m in a video game."
All together now: the most trustworthy covid-19 model is an ensemble - "Combining a multitude of predictions and projections, modeling teams hone the uncertainty. Earlier this spring, a paper studying covid forecasting appeared on the medRxiv preprint server with an authors’ list running 256 names long."
Anthropic is the new AI research outfit from OpenAI’s Dario Amodei, and it has $124M to burn - "As AI has grown from a menagerie of research projects to include a handful of titanic, industry-powering models like GPT-3, there is a need for the sector to evolve — or so thinks Dario Amodei, former VP of research at OpenAI, who struck out on his own to create a new company a few months ago."
Concerns & Hype
The hidden work created by artificial intelligence programs - "What’s keeping East Asians from leadership in US firms? As AI programs become more prevalent, researchers say that companies need to prioritize behind-the scenes workers and those affected by the programs."
'AI' is being used to profile people from their head vibrations – but is there enough evidence to support it? - "Digital video surveillance systems can’t just identify who someone is. They can also work out how someone is feeling and what kind of personality they have. They can even tell how they might behave in the future."
Report finds startling disinterest in ethical, responsible use of AI among business leaders - "The survey, conducted in February and March, features the insights of 100 AI-focused leaders from the financial services sector, with 20 executives hailing from the US, Latin America, Europe, the Middle East and Africa and the Asia Pacific regions. "
65% of execs can’t explain how their AI models make decisions, survey finds - "Elevate your enterprise data technology and strategy at Transform 2021. Despite increasing demand for and use of AI tools, 65% of companies can’t explain how AI model decisions or predictions are made."
Dumbed Down AI Rhetoric Harms Everyone - "When the European Union Commission released its regulatory proposal on artificial intelligence last month, much of the US policy community celebrated."
This $5 billion insurance company likes to talk up its AI. Now it's in a mess over it - "A key part of insurance company Lemonade's pitch to investors and customers is its ability to disrupt the normally staid insurance industry with artificial intelligence."
That's all for this week! If you are not subscribed and liked this, feel free to subscribe below!
Copyright © 2021 Skynet Today, All rights reserved.