Musicians have long criticized streaming services for their abysmal revenue sharing programs. In 2021, for example, as much as 97 percent of Spotify's over 6 million listed artists earned less than 1,000. Last year, the company announced a new system offering fractions of a cent per track, all of which is now based on even more stringent rules. But there was apparently a way to earn some real dividends from those songs--provided you have access to thousands of bots, hundreds of thousands of AI-generated songs, and are willing to risk receiving a federal grand jury indictment for wire fraud and money laundering. That's what a man named Michael Smith in North Carolina is currently facing, according to a DOJ announcement on September 4. Unsealed filings from US prosecutors accuse Smith of scamming digital streaming platforms including Spotify, Apple Music, Amazon Music, and YouTube Music of over 10 million in royalty payouts between 2017 and 2024.

This paper aims to advance our understanding of how Visual Language Models (VLMs) handle privacy-sensitive information, a crucial concern as these technologies become integral to everyday life. To this end, we introduce a new benchmark PrivBench, which contains images from 8 sensitive categories such as passports, or fingerprints. We evaluate 10 state-of-the-art VLMs on this benchmark and observe a generally limited understanding of privacy, highlighting a significant area for model improvement. Based on this we introduce PrivTune, a new instruction-tuning dataset aimed at equipping VLMs with knowledge about visual privacy. By tuning two pretrained VLMs, TinyLLaVa and MiniGPT-v2, on this small dataset, we achieve strong gains in their ability to recognize sensitive content, outperforming even GPT4-V. At the same time, we show that privacy-tuning only minimally affects the VLMs performance on standard benchmarks such as VQA. Overall, this paper lays out a crucial challenge for making VLMs effective in handling real-world data safely and provides a simple recipe that takes the first step towards building privacy-aware VLMs.

For instance, Philadelphia-based fintech firm cred.ai, The card was licenced by payments network Visa and issued by Wilmington Savings Fund Society, FSB. The credit optimizer tool uses an AI algorithm to improve the user's debt-to-credit ratio, which accounts for up to 30% of a FICO score that evaluates a person's creditworthiness in the US. Apple, too, uses AI to determine a user's credit limit on the Apple Card. Closer home, Gurugram-based fintech firm OneBanc has developed a card to connect various banking systems.

When the sale of Heather Lord's house completed, her conveyancing solicitor attempted to transfer funds to her bank account. An automated message warned him that her name did not match the name on the account he was paying into, and that the transfer would therefore be at his own risk. "I phoned the bank and asked who I was," she says. "They told me I was Ms Heather Audrey Lord. This was also refused by the system! As we didn't want to risk my money disappearing, I asked for a cheque, made out to Heather Lord, which was processed with no problem."

A sign of Google is seen at Google's stand during the annual meeting of the World Economic Forum ... [ ] (WEF) in Davos, on January 21, 2020. In an article titled Amazon's Impending Invasion Of Banking, I wrote: "Amazon has no incentive to cut banks out of the lending or deposit business. Amazon can make more money by providing technology services to help financial institutions underwrite, process, and service loans. Banks will gladly pay for this, because Amazon will do it for a lower cost that what banks incur to do it today." My argument then, as it is now, is that Amazon is poised to be a vendor--not a competitor--to financial institutions.

Machine learning is the latest to make waves in the field of Information Security, and for good reason. The support of complex algorithms that'learn' and grow is invaluable to human analysts, allowing them to focus on larger tactical fights and strengthen security systems to be virtually bulletproof. In both routine and structural changes to Information Security, machine learning plays an increasingly important role and will continue to do so, leading into the coming years. InfoSec refers to the systems, tools and processes that are designed and then deployed to field sensitive and confidential data from being compromised or tampered with. Disruption, modification and destruction of data are some of the more common results of InfoSec breaches.

The implementation of machine vision in finance is proving to be beneficial not only to businesses but also to consumers in general. Although it seemed far-fetched five years ago, financial institutions have been using advanced technologies in their operations. These technologies enable organizations to improve the customer experience and lighten employees' workloads. The global machine vision market is expected to grow at a CAGR of 7.7% to reach USD 18.24 billion by 2025. There, however, remains a certain level of skepticism amongst the institutes and consumers in the adoption of machine vision in the finance sector.

The iOS 11.2 update brought the much-awaited Apple Pay Cash features to iPhone and iPad users this week. The Apple Pay improvements were announced during Apple's Worldwide Developers Conference in June. Apple announced on Tuesday the features have rolled out to all users in the United States. "Apple customers already use iMessage to communicate with friends and family, and now they can get paid right within those everyday conversations, or by just asking Siri to pay someone," Apple said in its announcement. "They can use the debit and credit cards they already have added to Apple Pay, so there's no need to install an app or create yet another account."

The advantages of analytics to customer service have already been shown. Now the question becomes: How can analytics be used to improve security? Organizations are collecting more and more data. And while rich data allows personalized service, detailed data about real people (rightly) often raises concerns. Just as this data is increasingly valuable to organizations, it can be valuable to criminals as well, leading to an ever-escalating series of data breaches.

This "Weekend Think" article is adapted from Brett King's latest book, Augmented: Life in the Smart Lane, published last year. Since 2005 I've been predicting the fundamental decline of branch banking. For almost 10 years I fought bankers who decried my assessment that branches would cease to be the most important channel in banking, to be replaced by far more efficient mechanisms for revenue generation and relationship. Today the discussion is increasingly resorting to a sort of desperate plea: "But branches aren't going to die completely, are they?" No one who is watching these trends today is still saying branches will grow.