- Metaproof AI
- Posts
- ChatGPT creator itself launches AI detector tool
ChatGPT creator itself launches AI detector tool
DeepMind AI’s human speed, AI’s accurate global warming forecast, plus more top stories
Metaproof Team
February 02, 2023
👩💻 Metaproof AI
Hey frens, this is Sophia from 👩💻 Metaproof AI, the weekly newsletter where we keep you informed on how artificial intelligence (AI) is shaking things up across industries.
Stay informed via our Twitter and Subscribe here to get your AI news every Thursday.
📲 By the numbers
AI in the retail space is facing rosy prospects (Source):
AI adoption in retail and consumer products is expected to increase from 38% in 2022 to more than 75% by 2028, according to SkyQuest.
The machine learning segment registered 23.7% CAGR in 2022 and is anticipated to cross a 27% growth rate by 2028.
Customer relationship management (CRM) in retail had a 36% revenue share and is predicted to cross 43% share in 2028. Retailers can promote client loyalty and solid relationships by utilizing chatbots, search engines, and other technologies like AI, which will increasingly focus on planning and product recommendations.
📫 News & trends
OpenAI has launched a new tool aimed at helping teachers detect if text was written by a student or by an artificial intelligence program. The tool is a response to growing concerns that the use of AI technology in academic settings could fuel academic dishonesty and hinder learning. The details so far:
ChatGPT has been banned in public schools in several states in Australia, and as a result OpenAI has been working to craft new guidelines for educators, part of which is this new tool to detect AI-written text.
OpenAI emphasized the limitations of its detection tool, but said that besides deterring plagiarism it could help to detect automated disinformation campaigns and other misuse of AI to mimic humans.
The longer a passage of text, the better the tool is at detecting if an AI or human wrote something.
While some may view the concern over AI-written text as "overblown," it is still important to take steps to address academic integrity. Educators are urged to continue to look for ways to mitigate the risks associated with AI in the classroom and promote the responsible use of technology in academic settings.
DeepMind's latest AI, AdA, has achieved human-level accuracy and speed in solving tasks in a 3D virtual world. AdA's ability to perform tasks involving navigation, planning, and object manipulation suggests progress toward a “generally intelligent AI.”
While AdA's breakthrough is significant, AI still requires specific training and is not yet capable of functioning at a human level in the real world. AI's reliance on specific training limits its potential to perform a wide range of tasks, unlike humans. In addition, AdA's success in a controlled virtual environment is yet to translate to real-world performance and safety.
At any rate, the development is viewed as a possible step toward generally intelligent AI that could master an array of jobs in the real world.
The world is on the brink of breaching a critical climate threshold, with the likelihood of exceeding 1.5C of warming over industrial levels in the next decade and a 50% chance of surpassing 2C warming by mid-century, according to a new study.
Thanks to AI, the study predicted that there is nearly a 70% chance that the two-degree threshold will be crossed between 2044 and 2065. The ramifications are grave: Even half a degree of warming would lead to a 10-fold increase in the number of ice-free summers in the Arctic, as well as double the amount of lost habitat for plants and three times the amount for insects.
Here’s how AI worked n the research:
The scientists utilized a neural network, or a type of AI that recognizes relationships in vast sets of data, to train the system to analyze a wide array of global climate model simulations.
They asked the AI to determine timelines for given temperature thresholds.
To check the AI’s prediction prowess, they entered historical measurements and asked the system to evaluate current levels of heating already noted.
Using data from 1980 to 2021, the AI passed the test, correctly observing both the 1.1C warming reached by 2022 and the patterns and pace observed in recent decades.
In December 2022, OpenAI introduced ChatGPT, an AI language model capable of generating human-like text based on input. The potential applications for ChatGPT are diverse, ranging from messaging people on dating apps to writing essays. However, the use of ChatGPT raises ethical concerns. The AI model's ability to generate convincing text has the potential to spread misinformation and perpetuate harmful beliefs.
It's worth noting that not all chatbots use AI like ChatGPT. And while ChatGPT isn’t alive, it IS intelligent. It has yet to undergo a thorough evaluation using the Turing test to determine its level of human-like intelligence, but researchers believe it has passed the test nonetheless.
How to use ChatGPT:
Open chat.openai.com (Opens in a new window) in your browser.
Create an account and then, boom, you're plugged in. The tool is completely free to use, although there are rumors around a more powerful paid version in the future.
The United States has released a voluntary set of recommendations aimed at making companies more responsible in their development of AI systems. The guidelines follow the European Union's recently introduced Artificial Intelligence Act, which sets out a framework for the ethical development of AI systems. What the US guidelines recommend in a nutshell:
Consult domain experts, users, and affected communities when assessing the impact of AI systems, helping ensure the systems are designed with the end user in mind and that any potential negative consequences are considered before deployment.
Be diverse in terms of disciplinary background, expertise, and experience when developing AI systems, ensuring that a wide range of perspectives are considered when designing the systems and that any biases are minimized.
Just how rigorously big tech companies will adopt the guidelines remains to be seen, and it’s worth noting there’s still no way for third parties to verify if the guidelines are being followed, potentially limiting their effectiveness.
The demand for AI-focused cryptocurrency projects has skyrocketed in recent months, as interest in AI continues to grow. According to Google Trends, the worldwide score for the search term "AI" was an impressive 94 out of 100 during the week of January 22-28, 2023 - the demand partly driven by major tech investments, such as Microsoft's billions of dollars invested into OpenAI.
As a result, projects such as Fetch.ai, Singularitynet, and Ocean Protocol have seen their native tokens increase significantly in value over the past month. The demand for AI-based crypto assets and the popularity of these tokens may not last, though, and it’s worth noting that AI-related crypto assets are currently ranked below the top 75 in terms of market capitalization.
While the surge in demand for AI-focused cryptocurrency projects drums up excitement, it is important to approach these investments with caution and to consider the long-term potential of these assets, performing research and due diligence before committing to it.
AI and machine learning have the potential to transform the financial services industry and provide greater access to underserved populations. AI can do the following and more:
Streamline the lending life cycle by reducing processing time for loan applications and providing proactive customer support
Identify risk profiles and assess the ability to pay outside of the traditional target credit score
However, it's crucial that AI be used correctly to avoid biased decisions as well as tailor it to the unique needs of subprime customers. Despite these challenges, the use of AI in the financial services sector shows great promise for improving access and reducing friction for those seeking optimum financial solutions.
As companies continue to search for ways to streamline and improve their hiring process, many are turning to AI to write better job postings, with potential benefits including improved accuracy, reduced time and effort, and the ability to create job postings in multiple languages. Yet while AI-generated job postings have been shown to attract a more diverse pool of applicants, there are concerns about the limitations of AI in writing job postings, including the lack of human creativity and the potential for algorithmic bias.
While AI has the potential to revolutionize the hiring process, it is critical to limit its use as a tool, not as a replacement for human judgment. The limitations of AI, including algorithmic bias, must be taken into consideration. Weighing the benefits and limitations of AI, alongside using it in conjunction with human judgment, will go a long way.
Can the brain, with its limited realization of precise mathematical operations, compete with advanced AI systems implemented on fast and parallel computers? As AI systems continue to advance, it's worth examining the potential of the brain to compete with these systems.
Historically, AI has been derived from human brain dynamics. However, as compared to deep learning, brain learning has a number of important limitations (DL); for instance, efficient DL wire topologies (architectures) have several tens of feedforward (consecutive) layers, whereas brain dynamics have just a few feedforward layers. The race for efficient intelligence continues, and the potential of the brain to inspire the development of more advanced AI systems is worth exploring.
AI is revolutionizing the online slot gaming experience, making gameplay more immersive and personalized. AI can improve graphics and audio quality, providing a more lifelike and engaging experience for players. It can introduce new twists and turns to slot games, making them more interesting and dynamic. Machine learning can also be used to create a more personalized experience for players, where the level of engagement and enjoyment is barely seen previously.
Its potential limitations in the gaming world exist, though, including some players finding AI difficult to understand and not fully appreciating the ways in which it is affecting their gaming experience. Additionally, AI can be used to manipulate players, leading to less enjoyable and potentially harmful gaming time. We'll watch how developments play out for both players and gaming operators.
An AI-powered solution is emerging to help patients and clinicians predict whether symptoms are due to COVID-19, influenza, or RSV. The research, led by Professor Farrokh Alemi and Professor Janusz Wojtusiak at George Mason University, highlights that among the key benefits of AI in this context is improved clinical triage and test-to-treat decisions. AI-powered symptom screening can help individuals understand how to proceed if their symptoms and home test results do not agree, increasing accuracy and reducing the risk of incorrect treatment.
Caution abounds, however, when one considers that at-home tests are not always accurate and require clinical review, and AI symptom screening may not be accurate for vaccinated individuals. Yet as the fight against the global health concern continues, the role of AI in healthcare systems remains an exciting - if not lifesaving - development.
The discovery of a lost work by Spanish playwright Lope de Vega is making waves in the world of literature and AI. AI-powered text recognition tools were used to identify the author of the play, with traditional analysis confirming the AI's suggestion that Lope de Vega was the author. Lope de Vega is considered Spain's second most important author, after Miguel de Cervantes, with prolific writing that involved around 3,000 sonnets, three novels, and some 500 plays.
The play was discovered in the archives of the National Library of Spain, which purchased the work from a duke in 1882. The use of AI-powered text recognition tools allowed researchers to identify the author of the play much faster than traditional methods would have allowed, making it a significant breakthrough in the field of literary research.
The highly anticipated film Here, starring Tom Hanks and Robin Wright and with a planned 2024 release, is set to harness cutting-edge AI de-aging technology in its production. Directed by Robert Zemeckis, the film will feature the technology of AI firm Metaphysic, the company behind all those eerily accurate Tom Cruise and Bill Hader deep fakes from the past few years.
While still in development and testing, the use of AI de-aging technology has floated worries that it could lead to the erasure of important historical events and figures, while others believe it has the potential to revolutionize the way we tell stories on screen.
As AI continues to surge, it's becoming increasingly clear that the current cloud-computing infrastructure won't be enough to support its demands. In order to accommodate the rise of AI, a new framework is needed to support its digital infrastructure.
A few details at a glance:
One of the biggest challenges facing the adoption of AI is its high energy consumption. Data centers will need to be located near gigawatts of renewable energy sources and be equipped with new cooling systems and backup power to support AI's computational demands.
Cloud computing providers play a role in this "proxy war" for AI supremacy, offering powerful data-crunching capabilities to AI startups and positioning themselves as key players in the race for AI dominance.
As a counter to calls for the need for a specialized AI cloud infrastructure, naysayers say AI platforms may not be able to match the complexity of human brains, while others argue that the cost of upgrading data centers to support AI may be too high.
The need for a specialized AI cloud-computing framework appears to become more pressing, and companies that are able to adapt are seen to have a competitive advantage in this rapidly evolving landscape.
💸 Finance buzz
The $102M AI Powered Equity ETF (AIEQ), issued by ETF Managers Group and launched in 2017, leverages the power of IBM’s Watson supercomputer to select its portfolio holdings. It’s up 13.5% year to date as of Jan. 27, according to Morningstar data. (Source)
US investors, including the investment arms of Intel Corp and Qualcomm, accounted for nearly one-fifth of investments in Chinese AI companies from 2015 to 2021, a report released by tech policy group CSET showed. (Source)
🗣️ Quote of the week
"No, I am not alive. I am an artificial intelligence language model developed by OpenAI, I do not have consciousness or feelings. I am just a computer program designed to respond to text inputs and generate outputs based on patterns in the data I was trained on."
--
All content on this newsletter is for informational purposes only and does not aim to serve as or replace expert investment advice.
If you are a startup building in the metaverse / web3 ecosystem and are raising capital, please reach out to Sfermion. Sfermion is an investment firm focused on accelerating the emergence of the metaverse.
Stay informed via our Twitter and Subscribe here to get your AI news every Thursday.