Understanding the Invisible AI Mechanisms at Play in Personalizing Your News Feed, Stories, and Content Recommendations on Facebook and Instagram


How AI Influences What You See on Facebook and Instagram

Facebook and Instagram use artificial intelligence (AI) to personalize your content. Based on choices users make, the AI recommends posts that it predicts they’ll find interesting. Meta, the company behind these platforms, is striving for more transparency about how these algorithms work.

To shed light on this, Meta is sharing information about its AI systems that rank content on Facebook and Instagram. This includes clearer control options for users and in-depth details for experts to examine the systems. This openness is a part of Meta’s bigger commitment to transparency, especially with AI tech rapidly advancing.

Meta is releasing over 1,000 AI models, libraries, and datasets for researchers. They’re also launching a new suite of tools – Meta Content Library and API – that will provide comprehensive access to publicly available content on Facebook and Instagram. By introducing these tools to researchers early in the development process, Meta aims to receive feedback to build more user-friendly and transparent systems.


Microsoft brings new AI-powered shopping tools to Bing and Edge

Microsoft revealed some new AI-backed shopping tools for its Bing search engine and Bing AI chatbot in the Edge browser sidebar. This fresh set of tools appears promising, despite previous shopping features not being so popular.

The key feature involves Bing’s AI, based on the GPT model, creating shopping guides automatically. For example, if you search “college supplies,” Bing will generate a guide comprising similar items, compare their features, and tell you where to buy them. Interestingly, Microsoft will earn an affiliate fee if you purchase these items.

The implications of these AI-generated guides could affect websites specializing in shopping guides. It’s likely other search engines such as Google may soon offer similar tools. However, this might cause problems for high-quality, editorial shopping content creators.


Windows 11’s AI-powered Copilot enters public preview

Microsoft’s Windows 11 is getting an upgrade with Copilot, a new feature that brings AI technology straight to your fingertips. A test version of this feature will be available in the latest Windows Insider Preview for tech buffs willing to experiment. Once the update is installed, hitting the Windows key + C opens Copilot, a handy little helper right on your screen. It chats, keeps you updated, and can even make AI images with the same tech used in Bing Image Creator.

What’s more, Copilot can change certain Windows settings and carry out commands, potentially saving users the headache of navigating complicated settings. Examples include switching to dark mode, activating ‘Do Not Disturb’, and taking screenshots. Just a heads up, Microsoft is also using Copilot to show you ads they reckon you might find interesting.

As with all fresh updates, there’s a bunch of other improvements coming alongside Copilot. The Settings app gets a facelift with a new homepage featuring ‘cards’ – quick access to info or settings changes. Of the seven new cards, four relate to Microsoft account services like Microsoft 365 and Xbox subscriptions. The other cards show connected Bluetooth devices, let you quickly switch your desktop theme, and suggest settings tweaks.


Microsoft and Nvidia join $1.3bn fundraising for Inflection AI

Inflection AI, a new artificial intelligence company founded by former DeepMind co-founder Mustafa Suleyman, just reeled in a whopping $1.3 billion in funding. The money comes from tech bigshots like Microsoft and Nvidia, who are betting big on AI that can generate human-like text.

Suleyman and his team of 35, poached from top companies like Google and OpenAI, have cooked up a chatbot called Pi. It’s part of a wave of similar AI-powered chat products that have hit the market recently. And with Nvidia’s support, Inflection now has access to a whole heap of powerful computer processors, a vital ingredient in developing advanced AI.

Microsoft, who’s also backing Suleyman’s former venture, OpenAI, sees companies like Inflection as leaders in shaping the future of AI. They’re creating AI that’s not only revolutionary but also user-friendly.

Inflection’s chatbot, Pi, is designed for casual chit-chat with users through an app or social media platforms. Suleyman characterizes Pi as a friendly ear, rather than an information provider. This, he says, makes Pi easier to manage and safe for users.


Salesforce to invest $4 billion in UK on AI innovation

Salesforce, a US software company, has announced a $4 billion investment in its UK operations over the next five years, banking on strong demand for digital overhauls and artificial intelligence (AI). This follows a previous $2.5 billion input from 2018. This move is viewed as a major boost for the UK’s tech sector, particularly following recent criticism from Microsoft regarding regulatory decisions.

UK Prime Minister Rishi Sunak hailed Salesforce’s decision as a strong vote of confidence in the country’s economy. During his visit to Washington earlier this month, Sunak revealed plans for an international summit on AI safety in Britain later this year, as governments globally are attempting to create guidelines for this emerging technology.

Salesforce committed to cooperating with the UK government to spearhead the next wave of digital transformation in this new AI era. The firm emphasised the need for a clear regulatory framework that promotes innovation and safe, responsible use of AI. Despite disappointing quarterly results due to an uncertain US economy and reduced demand from financial services and tech firms, Salesforce is dedicated to delivering secure, trusted, and enterprise-ready generative AI to UK businesses.


Zenarate, an AI-powered agent coaching platform, raises $15M

Brian Tuite, ex-manager of a large contact center, teamed up with Rab Govil to found Zenarate, an AI-based training tool designed to prep customer service agents for real-life interactions. Think of it as a flight simulator, but for call center reps. The platform provides agents with feedback on their performance, highlights areas of improvement, and doesn’t need much tech set-up. The cherry on top? All training data is made-up, so no personal info is involved.

While Zenarate claims their AI isn’t biased, there isn’t much detail about how they prevent it, which could be a big issue. Nonetheless, investors seem keen, as Zenarate just raised $15 million in a funding round. Sean Cantwell from Volition Capital, the main investor, was impressed by Zenarate’s financials, highlighting its potential to shake up traditional training methods.

Zenarate, based in Palo Alto, will use the funds to grow their team. They’ve raised a total of $18 million so far and have a solid customer base that includes over a million users across 80+ companies.


New Clippy app gives us a taste of AI in Windows 11

An independent developer has breathed new life into Clippy, the once-familiar assistant from Microsoft Office of the late 90s and early 2000s, with a fresh app for Windows 11. This “new” Clippy, created by Firecube Studios, harnesses the power of OpenAI’s GPT-3.5 model to function like a smart assistant right on your desktop.

You can engage with Clippy for all kinds of help, even beyond Windows-related questions. Need to change a Windows setting? Ask Clippy. Wondering why there are several Penn Stations in the U.S? Clippy will try to answer that too.

In its current version, the user must supply their own OpenAI API key to use Clippy, but the final version will simplify setup. You can even control the length of Clippy’s responses by tweaking the number of API tokens used per question in the app’s settings.


Oracle adds generative AI to its human resources software

Oracle Corp announced it’s adding an artificial intelligence (AI) function to its human resources (HR) software. The goal is to help with tasks like creating job descriptions and employee performance objectives. AI systems that can mimic human responses have become quite popular in tech. Companies like Microsoft and Google are already using these systems in their search engines.

Oracle’s HR software is a tool used by many large companies for tasks like hiring new workers and doing performance reviews. Oracle’s new addition to this software is an AI assistant – a button that generates draft text for things like job ads and performance objectives. The idea behind a button rather than a chatbot is to ensure the outcome is reliable and safe.

Oracle’s aim isn’t for the AI to do all your work, but to give you a head start, providing useful info to kick things off. Oracle plans to use AI for more complicated HR tasks, such as writing job requirement listings that comply with local rules in different markets. This could turn decision-making times from weeks to hours, a real game-changer.


RelationalAI and Snowflake join forces to revolutionize enterprise AI decision-making

California AI startup RelationalAI is teaming up with cloud data warehouse provider Snowflake, introducing an AI “coprocessor” that merges relational knowledge graphs and complex AI abilities into Snowflake’s platform. Announced at Snowflake Summit 2023, this addition supports Snowflake’s bid to be a complete platform for business AI, and aligns with RelationalAI’s vision for smarter applications.

RelationalAI CEO, Molham Aref, explains that the integration simplifies data understanding for both humans and language models, comparable to how a knowledge graph works. This coprocessor lets Snowflake users carry out tasks like running knowledge graphs and predictive analytics without having to move their data elsewhere. It lets them build AI-focused applications such as fraud detection or supply chain optimization straight within Snowflake.

Running safely in the Data Cloud with Snowflake’s newly announced Snowpark Container Services, the AI coprocessor has seen early use in sectors like finance, retail, and telecommunications. Aref emphasizes how language models can access company data and translate queries into SQL commands to provide specific answers to broad questions.


Reka emerges from stealth to build custom AI models for the enterprise

Reka, a startup founded by AI experts from DeepMind, Google, Baidu and Meta, came out of stealth mode today with a cool $58 million in funding. The brains behind it want to create customized AI models to help companies tackle specific tasks. While big AI models like OpenAI’s GPT-4 are good at analyzing and generating text, they’re not so hot when it comes to more precise tasks, like writing in a brand’s specific style.

Reka’s first product, Yasa, is an AI assistant that understands images, videos and data as well as text. Although it’s similar to models like GPT-4, Yasa has a special trick – it can be tailored to a company’s specific data and applications. This feature can help companies take advantage of AI technology without needing a bunch of expert AI engineers in-house.

But that’s just the start. Next up, Reka wants to create AI that can handle even more types of data and update itself without needing retraining. For select customers, they also offer a service to adapt their AI models to the company’s unique data.


Cyera raises $100M to expand data security platform for AI-driven enterprises

Cyera, a data security firm, recently announced it scored a hefty $100 million in series B funding, with Accel at the helm and other contributors such as Sequoia, Cyberstarts, and Redpoint Ventures. Since kicking off in 2022, Cyera’s raked in $160 million in total and witnessed its revenue skyrocket 800% year-over-year. That’s pretty impressive, given the tough economy and tight funding landscape!

Cyera uses artificial intelligence (AI) to boost data security in businesses, especially those heavy hitters in the S&P 500 that use a mix of cloud services. Traditional companies struggle to keep up with data security needs because their old systems just don’t cut it anymore. But Cyera is flipping the script with its platform that helps businesses really get to grips with their data: what it is, why it’s important, and how much it’s worth.

With the latest cash injection, Cyera plans to turbocharge the development of its cloud-native platform. The aim is to help security teams handle data security incidents, manage rules and controls, and streamline operations across all their data.


OpenAI brings the competition to DeepMind’s doorstep with new London office

OpenAI, a prominent AI startup backed by Microsoft, plans to establish its first international branch in London. This new location will focus on expanding OpenAI’s research and engineering capabilities, while engaging with local communities and policymakers. CEO Sam Altman sees this move as a chance to attract top-notch talent and push forward AGI (Artificial General Intelligence) development and policy.

London is a notable choice since it is not only the home to DeepMind, Google’s main AI research division, but also a hub of data science talent and a booming center for AI startups. As of 2021, the city housed over 1,300 AI companies and led the U.K. in venture capital investments.

The choice also reflects OpenAI’s political interests, as the company wants to convince U.K. authorities to apply light-touch regulation on AI. Recently, Altman made a public call for balanced AI regulation and highlighted the threat of deepfake disinformation. While OpenAI will operate within the rules of the EU’s AI Act, Altman hinted that it might withdraw from the European Union if it cannot meet these regulations.


Microsoft president goes to Europe to shape AI regulation debate

Microsoft’s President, Brad Smith, traveled to Brussels to discuss the future of artificial intelligence (AI) regulations. Smith has been vocal about this issue, teaming up with tech bigwigs like Elon Musk, CEO of Twitter and Tesla, to call for rules around AI use. AI has been a hot topic due to innovations like Microsoft’s OpenAI’s ChatGPT, a cutting-edge system that’s sparking a lot of interest.

This comes at a time when the European Union (EU) is refining a set of regulations, referred to as the AI Act. These regulations could potentially be a yardstick for other countries. Smith, in his blogpost, voiced his intention of contributing to this critical discourse.

According to Smith, Microsoft’s own plan for AI oversight, which involves steps like setting up safety measures for AI systems in charge of crucial infrastructure, aligns with the EU’s upcoming laws. Moreover, Smith encouraged cooperation on AI governance between the EU, the US, G7 nations, India, and Indonesia. The goal? To create AI rules that reflect their common values and principles.


AI will change software development in massive ways, says MongoDB CTO

Artificial Intelligence (AI), particularly generative AI like ChatGPT from OpenAI, is set to massively improve the productivity of software developers, says Mark Porter, Chief Technology Officer of MongoDB. He disagrees with the notion that AI will render developers obsolete. Instead, he emphasizes how generative AI can assist in code creation, bug detection, and documentation lookup, making the coding process faster and more efficient.

Porter believes AI has the potential to transform the way developers write code and accelerate the creation of more applications. He feels that generative AI will drastically reduce the time it takes to write software and improve its quality. At a recent conference, MongoDB showcased new AI capabilities in their database, which many companies utilize for their AI applications.

One novel functionality Porter highlighted is the inclusion of vector values as a native data type in MongoDB’s database. This allows developers to store and retrieve context vectors from large language models, leading to more precise query answers. He also mentioned how AI and databases can work together, with AI providing functions like article summarization.


A.I. May Someday Work Medical Miracles. For Now, It Helps Do Paperwork

In the field of healthcare, AI is acting more like a paper-pusher than a mastermind, helping to ease the burden of digital paperwork. Dr. Matthew Hitchcock, a family physician in Chattanooga, uses an AI tool that records patient visits, summarizes them for treatment and billing, and helps him complete his daily documentation in roughly 20 minutes. This use of AI is a far cry from his previous after-hours documentation grind.

One AI company, Abridge, has developed software that records and summarizes doctor-patient conversations, translating medical jargon into simple language. Patients, in turn, receive a summary of their visit that they can refer back to, helping them better understand and remember their healthcare instructions.

For now, doctors are grateful to have AI take on the role of a diligent scribe, freeing their evenings from the drudgery of digital paperwork. While there are high hopes for AI’s future contributions to healthcare, the technology won’t be solving staffing shortages any time soon.


Analysts see A.I. as a double-edged sword for 3 big tech stocks

Big Tech is diving deeper into AI and Wall Street is taking note. Analysts suggest this could push tech giants like Alphabet, Amazon, and Meta to new heights – but it might not be all smooth sailing.

For Alphabet, there’s some worry. UBS analysts fear AI could mess with Google’s ad-supported search results, posing a risk to revenues. However, they still see potential for AI to unlock new ways for the company to make money.

On the Meta front, the mood is more upbeat. UBS analysts believe Meta’s use of AI could be a game-changer. They’re particularly excited about how AI could boost user engagement on Meta’s platforms, with chatbots predicted to make a big revenue impact. The fact that Meta is keeping costs in check is another plus.

As for Amazon, it’s all systems go according to Roth MKM. They think Amazon could ride the AI wave, particularly with expected growth in Amazon’s cloud unit, Amazon Web Services. Also, they believe Amazon’s operating margins are underrated and should get a boost from recent cost-cutting measures.


More People Are Going Blind. AI Can Help Fight It

Eye problems have become a major issue in the UK, with ophthalmology being the busiest medical specialty in the National Health Service. The main causes of blindness are diabetic eye disease and age-related macular degeneration (AMD). 

The number of people with eye problems has increased by over a third in the past five years. The challenge lies in diagnosing and treating patients in a timely manner, leading to delays and preventable blindness. 

Optical coherence tomography (OCT) devices are advanced eye scanners that can help, but optometrists often lack the expertise to analyze the scans. Artificial intelligence (AI) can bridge this gap by bringing expert analysis to the community. 

A collaboration between DeepMind and Moorfields Eye Hospital demonstrated that an AI system can analyze OCT scans as effectively as human ophthalmologists. The goal now is to validate the system and train it on diverse data to ensure it works for all patients. Once achieved, the AI system can be deployed at scale, identifying and prioritizing patients for treatment. This will alleviate the burden of chronic eye diseases like AMD. 

The future of ophthalmology AI is comparable to Thomas Edison’s invention of the light bulb—it requires a network of innovations and collaborations to make it widely accessible. With the integration of OCT machines, cloud connectivity, and national transformation programs, AI in ophthalmology is on the horizon.