Pervasive artificial intelligence: what happened in IT in early autumn

Date:

Share post:

Most startups are developing neural networks for various industries, fintech is developing AI solutions, and payment services are appearing on social networks. About events in the IT world – in the material “Xu Techs”

Glossary

Explainable Artificial Intelligence (XAI) is a branch of artificial intelligence that aims to make the decisions and actions of AI systems and AI models more understandable to people.

The digital currency of central banks (Central Bank Digital Currency, CBDC, or CBCB) is a digital analogue of national fiat currencies (traditional currencies, such as rubles, dollars, yuan), which are issued, regulated and guaranteed by central banks (CB).

Generative AI (GenAI) is a category of artificial intelligence (AI) algorithms that can generate new results based on the data on which they are trained.

A Large Language Model (LLM) is a deeply trained neural network used for natural language processing.

HNWI (High Net Worth Individuals) are high-income individuals whose capital ranges from $1 million to $30 million.

UHNWI (Ultra High Net Worth Individuals) – individuals with a net worth of at least $30 million.

The digest was prepared by the red_mad_robot analytical center.

Trends of the month: explainable AI, AI hype, and cross-platform from Apple

Explainable AI integrates into people’s lives and work

Explainable AI systems and models are increasingly being used in areas such as healthcare, banking, education, agribusiness, and industry, where the consequences of errors or biases can be serious or even fatal for a business. Therefore, XAI has become a more important area. Here are the latest industry trends.

Concept-based explanations. A way to “explain” AI decisions in high-level concepts such as emotions, goals, values, or categories. For example, a conceptual explanation of facial recognition systems might say that the AI ​​model identifies a person as happy because they have a smile on their face, raised eyebrows, and open eyes.

Counterfactual explanations. These are, as a rule, hypothetical reasoning, thought experiments. The actual result is compared with the alternative, and the factors that led to a particular decision are highlighted. For example, if we explain the possible refusal of an AI model to approve a loan, then the reasons may include the applicant’s insufficiently high income, large debt, or insufficiently good credit history. Moreover, it will indicate how much each factor influenced the AI ​​model’s decision making.

Interaction with the model. Users can interact with the AI ​​model to receive explanations for its behavior. For example, a person can find out why the AI ​​chose a particular treatment or medicine. This increases the user’s level of trust and also gives the user the opportunity to gain new knowledge by understanding the logic and assumptions behind the algorithms.

The startup market is experiencing AI hype

In September 2023, Y Combinator (YC) held a Demo Day. AI has completely taken over the startup market: more than 60% of startups presented at this accelerator season are related to artificial intelligence technologies. Startups are increasingly developing AI solutions for specific verticals and sectors. These sectors can be anything, the main thing is the huge market potential and its volume. For example, AI for banking, education, retail, etc.

Success, according to YC, is finding pain and problems in corporations or corporate business processes and solving them using a niche AI ​​service or AI model.

Apple will promote cross-platform

Apple does not mention artificial intelligence in its announcements, unlike Microsoft and Google, which put forward ambitious concepts for the development of AI. But that doesn’t mean Apple isn’t developing it. First, the company is developing an AI-based chatbot to compete with ChatGPT. Secondly, several AI features were presented at the Apple Wonderlust conference. For example, with the help of neural networks, the iPhone 15 has improved noise reduction, background blur for portrait photography, and predictive text. And in the latest Apple Watch Series 9 Neural Engine, thanks to machine learning, Siri accuracy will increase by 25%.

At the same time, the company has another direction – cross-platform in the overall ecosystem of products (the ability to work with several hardware platforms at once). In early September, Apple announced the App Store for the Vision Pro mixed reality headset. The project is almost ready for use: bigtech notes that the beta version of VisionOS from the App Store will be released before the end of autumn this year.

New AI kit: “money on autopilot”

One of the most important advances in generative AI is the ability to process and create not only text, but images and other forms of content. The use of this technology in the field of consumer financial services was analyzed by the a16z venture capital fund, predicting the emergence of “autopilots” that could bring to life the idea of ​​“self-propelled money.” In the future, the cost of applying for services will be virtually reduced to zero, which will allow the “autopilot” to constantly search the data set for the best conditions “without hands.”

This innovation, on the one hand, can democratize financial planning, on the other hand, create problems for financial institutions, such as a decrease in customer loyalty and an increase in the number of applications. Financial institutions will need new AI-powered tools to accurately assess transaction intent, combat fraud, and process applications more efficiently.

Events of the month

AI is taking fintech by storm. British fintech company Trade Ledger is implementing an AI bot based on Azure OpenAI (a cloud platform from Microsoft). The bot personalizes reports and offers clients suitable loan terms. Fintech company HighRadius has launched the FreedaGPT neural network, which combines the company’s developments and the OpenAI model: it answers questions in letters and can generate reports on financial requests. Software company Temenos has developed GenAI to automatically classify customers’ banking transactions.

Going beyond the social network . X (aka Twitter) is making moves in the payments space: the social network recently received a money transfer license in several states. WhatsApp in India now offers the ability to pay for goods and services.

Focus on ESG . Swedish fintech company Klarna has launched features that help users make “informed” purchases. The service allows you to compare prices from thousands of retailers and filter product search results by characteristics (color, size, customer ratings); The latest update adds a special filter to highlight conscious brands that have received a sustainability rating of 4 (good) or 5 (excellent) from Klarna partner Good On You.

Neuronews . According to business intelligence platform Crunchbase , in the first half of 2023, one in four dollars invested in US startups went to AI-related companies. Analysts from Innopolis University are confident: the volume of the AI ​​market in Russia has grown from 240 billion rubles. in 2019 to more than 600 billion rubles. in 2022. Leading positions in the development and implementation of AI are occupied by Yandex, VK and Sber. Here are some recent news from the global and domestic industry.

Yandex plans to adapt the YandexGPT neural network for the Middle East markets. The company has also resumed testing self-driving cars in the US, but under the Avride brand.
By the way, about the Middle East. The Abu Dhabi Institute for Technological Innovation has publicly released a large language model (LLM) of the Falcon 180B, which takes into account 180 billion parameters. It is superior in quality to open models and GPT-3.5. UAE-based technology company G42 is launching an AI model in Arabic. By the way, Microsoft is ready to cooperate with G42.

Microsoft is teaming up with digital pathology AI solutions company Paige to develop the largest artificial intelligence model for cancer detection.

To create a convenient system for contactless payment and personal identification, Amazon One trained a neural network based on millions of artificial palms.

Google has announced a slew of AI updates. For example, 20 new ready-made AI models optimized for enterprises have been added to the cloud service.

The kick sharing service Whoosh and the Neiry company are testing a scooter that is controlled using a neural interface.

Chinese scientists have demonstrated the SpiralE neural interface, which is presented in the form of a device made of a flexible polymer that transmits signals from the user to the computer. SpiralE is worn over the ear.

China Mobile has launched a 5G supernetwork with support for metaverses. The company also proposes to transfer the social credit system to them.

Hiber3D has introduced the ability to generate worlds for metaverses based on prompts.

Topic of the month: what will happen to fintech

According to a flagship report from consulting firm Boston Consulting Group, the fintech market will grow fivefold to reach $1.5 trillion by 2030, with the Asia-Pacific and North American regions becoming centers of gravity for startups and companies. These markets will be valued at $600 billion and $520 billion, respectively (more than 70% of the total market).

Digital payments (retail, crypto payments, b2b) and digital lending will dominate the market and generate about 60% of all revenue. At the same time, neobanking (digital banking), insurtech (insurance technologies), solutions for the development of financial infrastructure and wealthtech (technologies for wealth management) will also grow.

In general, investments are evenly distributed across regions, but there are differences in priority segments. For example, in developing regions, investors see prospects in payments and lending, and in developed regions – in financial infrastructure and wealthtech.

Several interesting trends by segment:

Digital payments . Instant payments will become the mainstay for many countries. CBDCs are gaining momentum, and cross-border payments are ripe for disruption—disruptive innovation.

Digital lending . GenAI, blockchain and ecosystems will become mainstream.
Wealthtech . So far, there is insufficient penetration of HNWI/UHNWI into digital. There will be a strong increase in interest in alternative assets (this is investing in any asset class, with the exception of stocks, bonds and cash) in all countries of the world.

Insurtech . The market will increasingly experience hyper-personalization of offerings due to the growth of consumer lifestyle and health data through wearable devices.
Neobanking . Neobanks for SMEs are becoming increasingly important to investors. The B2b model demonstrates the potential to create a scalable customer base that will allow you to generate stable profits.

LEAVE A REPLY

Please enter your comment!
Please enter your name here

Related articles

Delegate the Routine: Why Companies Should Automate Business Processes

Automation and digitalization are called key factors for a company's success. We figure out how this helps to...

The History of Boston Dynamics: From Creepy Combustion Engine Robots to Useful Electric Dogs

Boston Dynamics creates robots that work in warehouses, measure human temperature, and do somersaults better than acrobats. We...

The 4 best Javascript training courses in 2024

📢 Too busy to read it all? Here is the best Javascript training in 2024: After using Clic Compétences' JavaScript...

How to Keep Cyber ​​Threats Out of the Physical World and What Business Thinking Has to Do With It

Cyberattacks now pose not only an informational but also a physical threat. How to reduce such risks and...