Weekly Roundup: Cutting-Edge AI News and Top Stories for May 2023

We discuss the profit-driven hype cycle of AI misinformation and its potential consequences, the opportunities and challenges of AI for older adults, and the autonomous improvement of AI agents through negotiation games and feedback.
Weekly Roundup: Cutting-Edge AI News and Top Stories for May 2023

AI Misinformation: The Profit-Driven Hype Cycle and its Consequences

In the AI discourse, two narratives dominate: AI as a revolutionary tool and AI as an existential threat. This article reveals both as profit-driven misinformation strategies, creating urgency, distorting expectations, and glamorizing AI. The author warns that this not only distracts from real AI issues but also exploits the vulnerable, leading to a market saturated with individuals lacking fundamental AI understanding.

Highlighting the case of LinkedIn influencer Zain Kahn, the article exposes how such figures profit from AI hype, selling products and services based on inflated promises. Drawing parallels with the cryptocurrency scams of 2021-22, the author warns of the potentially severe consequences of AI misinformation, given the technology's societal integration. The piece concludes with a call for critical consumption of AI information to avoid falling prey to these narratives.

The Golden Age of AI: Opportunities and Challenges for Older Adults

The technology industry is witnessing a significant disruption with the advent of Large Language Model (LLM) AI, such as OpenAI's ChatGPT. This innovation has sparked a surge of investment and development, leading to the emergence of Conversational AI in the Gartner Hype Cycle. Amidst the media hype and skepticism, the healthcare industry, particularly age-related sectors, sees potential in AI for streamlining processes, augmenting labor shortages, and enhancing patient interaction. AI's role in improving staff management, fall avoidance, and customer-facing interactions is also gaining traction.

However, the rapid evolution of AI has also raised concerns about bias and ethical issues. Despite these challenges, the report suggests that we have entered a 'Golden Age of AI', with AI likely to become part of a 'wellbeing' infrastructure incorporated into home design and remodeling. The report, based on interviews with 26 AI experts, provides a comprehensive overview of the fast-evolving AI technology and its potential benefits for older adults. It serves as a valuable resource for investors, healthcare professionals, senior living organizations, and others interested in the intersection of AI and aging.

Autonomous Improvement of AI Agents through Negotiation Games and Feedback

Researchers are exploring whether multiple Large Language Models (LLMs) can autonomously improve each other through a negotiation game. The game involves two LLMs playing the roles of a buyer and a seller, with a third LLM acting as a critic providing feedback for strategy improvement. The study aims to determine if strong AI agents can be created with minimal human intervention. The findings reveal that only a subset of LLMs can self-play and improve from AI feedback, with weaker models failing to understand the game's rules or incorporate feedback for improvement. The study also found that models' abilities to learn from feedback differ when playing different roles, and stronger agents can consistently improve their performance using previous experiences and iterative AI feedback. However, this also increases the risk of breaking the deal.

The study, which uses different LLMs for different roles and the deal price as the evaluation metric, provides an insightful exploration of having models autonomously improve each other through game playing and AI feedback. The findings have profound implications, including the potential to create strong AI agents with minimal human intervention and the associated risks of limited human oversight. The research also serves as a meaningful initiative for improving language models' negotiation in a game setting using AI feedback.

Kraft Heinz Boosts Sales by $30M Through AI-Enhanced Supply Chain Visibility

Kraft Heinz has reported a $30 million increase in sales after integrating artificial intelligence (AI) into its supply chain visibility functions. The food manufacturer has utilized AI to streamline operational priorities, automate service risk identification, and reduce operator alerts by 42%. The company's service levels have consistently improved since September, with case fill rates in March rising by over five percentage points to just above 95%.

CEO Miguel Patricio highlighted a partnership with Microsoft and its Azure cloud service to digitize the supply chain, including a "control tower" for real-time visibility into plant operations and automated distribution across product categories. This initiative is one of Kraft Heinz's most significant technology investments.

Neural Volumetric Memory (NVM) Enhances Legged Robot Locomotion on Challenging Terrains

Researchers have proposed a neural volumetric memory architecture (NVM) to improve the locomotion of legged robots on difficult terrains. Unlike traditional visual-locomotion methods, which typically stack images channel-wise to address partial-observability, NVM explicitly accounts for the SE(3) equivariance of the 3D world.

The NVM aggregates feature volumes from multiple camera views by applying 3D translations and rotations to bring them back to the robot's ego-centric frame. This approach allows the robot to aggregate information from previous frames and accurately infer the occluded terrain underneath it, despite large and spurious changes in camera pose.

The researchers tested the learned visual-locomotion policy on a physical robot, demonstrating performance gains over previous methods on challenging terrains. They also showed that the representation stored in the neural volumetric memory captures sufficient geometric information to reconstruct the scene.

The team conducted real-world experiments in both indoor scenarios with stepping stones and in-the-wild scenarios to validate their method in different real-world scenes beyond the simulation. The results suggest that the NVM approach could significantly enhance the capabilities of legged robots in navigating complex environments.

FutureFinder: Revolutionizing Education and Career Pathways with AI-Powered Guidance

FutureFinder is an AI-powered platform designed to provide personalized guidance for individuals' educational and career pathways. The platform offers a range of features, including an AI Genie for customized advice, an in-depth assessment tool powered by GPT-4, an AI-powered essay analyzer, mock AI interviews, tailored recommendations, and detailed application guidance. It caters to high school students seeking college advice, college students exploring career paths, and working professionals considering further education or a career change.

The platform provides access to over 1,200 colleges and universities worldwide, offering personalized recommendations based on users' preferences and goals. It also offers in-app support from AI Genie and real advisors, a streamlined document management system, and a user-friendly interface. FutureFinder aims to provide comprehensive assistance at every stage of the academic journey, from exploring institutions to receiving offer letters, all free of cost.

AI's Role in Scriptwriting Fuels Hollywood Writers' Strike

The Hollywood writers' strike, initially triggered by compensation disputes in the age of streaming services, has now evolved to address the contentious issue of artificial intelligence (AI) in scriptwriting. The Writers Guild of America (WGA) is locked in negotiations with industry behemoths such as Disney, Netflix, and Warner Bros. Studios, advocating for a ban on AI technologies like ChatGPT in the generation of story ideas or scripts. The union's concern is that AI could undercut writers' compensation and credits by producing initial drafts of shows, thereby devaluing their craft and expertise.

In addition, the WGA proposes that scripts under its collective bargaining agreement should not be used to train AIs. This stance echoes a broader resistance against tech companies using online content to train AI models. However, the industry has rebuffed these proposals, offering instead to hold annual meetings to discuss technological changes. This strike underscores the escalating concern among workers across various industries about the potential of AI to automate or replace human jobs, raising critical questions about the future of work in the age of AI.

Trump Mocks CNN with Literal "Fake News" with a Deepfaked Video after Town Hall Event

Former President Donald Trump has shared a manipulated video clip parodying a CNN broadcast, in a move seen as a mockery of the network's reaction to his recent town hall event. The doctored video, initially posted by Trump's eldest son, Donald Trump Jr., features CNN anchor Anderson Cooper and was shared on Trump's Truth Social account. The clip has added fuel to the controversy surrounding CNN's decision to host the town hall, which saw Trump, a potential 2024 White House contender, promote false claims about the 2020 election and voice support for defendants charged in connection with the Jan. 6, 2021, Capitol attack.

Critics, including Democrats and media pundits, argue that CNN erred in providing a platform for Trump's controversial views. However, CNN's leadership has defended the decision, citing Trump's status as the Republican primary front-runner and the network's responsibility to challenge his false statements. Cooper, too, defended the network's choice during his show, acknowledging viewers' anger but emphasizing the importance of engaging with differing viewpoints.

ChatGPT Suffers Data Breach Amid Growing Concerns Over AI Safety

OpenAI's chatbot, ChatGPT, has been in the spotlight for both its capabilities and the concerns it raises regarding safety, privacy, and potential misuse. The AI's reputation took a hit recently when a bug in its source code led to a data breach, exposing sensitive user information. The breach affected approximately 1.2% of ChatGPT Plus subscribers active on March 20th, potentially compromising their payment information, including names, email addresses, payment addresses, credit card types, and the last four digits of credit card numbers. OpenAI has since patched the vulnerability and assured users of no ongoing risk to their data.

This incident comes amid growing apprehension about AI systems like ChatGPT. Geoffrey Hinton, known as the "godfather of AI", recently left Google to independently advocate for AI safety and control, highlighting the urgent need for investment in this area. In response to the breach, cybersecurity experts are urging individuals to employ reliable cybersecurity protection and to utilize platforms like Trend Micro's Free ID Protection to monitor their data and secure their online presence. As AI continues to evolve, the focus on safety and control becomes increasingly critical.

AI's Potential Threats to Human Health and Existence: A Call for Regulation and Safe Practices

Artificial intelligence (AI), while promising transformative solutions in healthcare, also presents significant threats to human health and well-being through social, political, economic, and security-related determinants. A recent analysis highlights three primary ways in which misused narrow AI can pose a threat: by increasing opportunities for control and manipulation of people, enhancing lethal weapon capacity, and rendering human labor increasingly obsolete. The study also discusses the potential existential threat posed by self-improving 'artificial general intelligence' (AGI), urging the need for effective regulation and a moratorium on the development of self-improving AGI.

The analysis underscores the potential misuse of AI in surveillance and manipulation of people, the development of lethal autonomous weapon systems, and the anticipated job losses due to AI-driven automation. The authors call on the medical and public health community to engage in evidence-based advocacy for safe AI, rooted in the precautionary principle. As AI continues to evolve rapidly, the focus on safety, control, and equitable distribution of benefits becomes increasingly critical to mitigate potential threats to human health and existence.

AI & Big Data Expo North America 2023: Delivering AI & Big Data for a Smarter Future

The AI & Big Data Expo North America, a conference and exhibition that took place this past week on May 17-18, 2023, at the Santa Clara Convention Center, CA, explored future of enterprise AI, edge AI, ethical AI, cybersecurity, and machine learning. The event featured presentations and panel discussions from industry leaders, including Amazon Web Services, Schlumberger, CVS Health, and SoftServe, among others. The conference provided insights into the latest advances in AI and big data, discuss the ethical implications of AI, and explore the potential of AI in various sectors, including healthcare and renewable energy.

The conference kicked off with a chairperson's welcome, followed by presentations on advances in edge AI, accelerating AI/ML innovation, AI-enabled healthcare, and the disruptive force of AI-driven insights. Panel discussions focused on ethical, responsible, and governance aspects of scaling AI safely, the convergence of AI with other disruptive technologies, and the present and future of AI development. The event also featured keynote presentations on generative AI's unexpected reality check and AI-driven insights as the most disruptive force ever to meet the worlds of data and business.

Learn about one of the stars of the show - eventualAi in our featured post here.

Artificial Intelligence News & Reviews

nextomoro is the most comprehensive source for Artificial Intelligence news & reviews. Learn how Artificial Intelligence will revolutionize the way we live and work.

Artificial Intelligence News & Reviews

Great! You’ve successfully signed up.

Welcome back! You've successfully signed in.

You've successfully subscribed to Artificial Intelligence News & Reviews.

Success! Check your email for magic link to sign-in.

Success! Your billing info has been updated.

Your billing was not updated.