Connect with us

NEWS

Privacy not a blocker for “meaningful” research access to platform data, says report

Published

on

privacy not a blocker for meaningful research access to platform data says report
368912557 2fc44d3709 b

European lawmakers are eyeing binding transparency requirements for Internet platforms in a Digital Services Act (DSA) due to be drafted by the end of the year. But the question of how to create governance structures that provide regulators and researchers with meaningful access to data so platforms can be held accountable for the content they’re amplifying is a complex one.

Platforms’ own efforts to open up their data troves to outside eyes have been chequered to say the least. Back in 2018, Facebook announced the Social Science One initiative, saying it would provide a select group of academics with access to about a petabyte’s worth of sharing data and metadata. But it took almost two years before researchers got access to any data.

“This was the most frustrating thing I’ve been involved in, in my life,” one of the involved researchers told Protocol earlier this year, after spending some 20 months negotiating with Facebook over exactly what it would release.

Facebook’s political Ad Archive API has similarly frustrated researchers. “Facebook makes it impossible to get a complete picture of all of the ads running on their platform (which is exactly the opposite of what they claim to be doing),” said Mozilla last year, accusing the tech giant of transparency-washing.

Facebook, meanwhile, points to European data protection regulations and privacy requirements attached to its business following interventions by the US’ FTC to justify painstaking progress around data access. But critics argue this is just a cynical shield against transparency and accountability. Plus of course none of these regulations stopped Facebook grabbing people’s data in the first place.

In January, Europe’s lead data protection regulator penned a preliminary opinion on data protection and research which warned against such shielding.

“Data protection obligations should not be misappropriated as a means for powerful players to escape transparency and accountability,” wrote EDPS Wojciech Wiewiorówski. “Researchers operating within ethical governance frameworks should therefore be able to access necessary API and other data, with a valid legal basis and subject to the principle of proportionality and appropriate safeguards.”

Advertisement

Nor is Facebook the sole offender here, of course. Google brands itself a ‘privacy champion’ on account of how tight a grip it keeps on access to user data, heavily mediating data it releases in areas where it claims ‘transparency’. While, for years, Twitter routinely disparaged third party studies which sought to understand how content flows across its platform — saying its API didn’t provide full access to all platform data and metadata so the research couldn’t show the full picture. Another convenient shield to eschew accountability.

More recently the company has made some encouraging noises to researchers, updating its dev policy to clarify rules, and offering up a COVID-related dataset — though the included tweets remains self selected. So Twitter’s mediating hand remains on the research tiller.

A new report by AlgorithmWatch seeks to grapple with the knotty problem of platforms evading accountability by mediating data access — suggesting some concrete steps to deliver transparency and bolster research, including by taking inspiration from how access to medical data is mediated, among other discussed governance structures.

The goal: “Meaningful” research access to platform data. (Or as the report title puts it: Operationalizing Research Access in Platform Governance: What to Learn from Other Industries?

“We have strict transparency rules to enable accountability and the public good in so many other sectors (food, transportation, consumer goods, finance, etc). We definitely need it for online platforms — especially in COVID-19 times, where we’re even more dependent on them for work, education, social interaction, news and media consumption,” co-author Jef Ausloos tells TechCrunch.

The report, which the authors are aiming at European Commission lawmakers as they ponder how to shape an effective platform governance framework, proposes mandatory data sharing frameworks with an independent EU-institution acting as an intermediary between disclosing corporations and data recipients.

Advertisement

It’s not the first time an online regulator has been mooted, of course — but the entity being suggested here is more tightly configured in terms of purpose than some of the other Internet overseers being proposed in Europe.

“Such an institution would maintain relevant access infrastructures including virtual secure operating environments, public databases, websites and forums. It would also play an important role in verifying and pre-processing corporate data in order to ensure it is suitable for disclosure,” they write in a report summary.

Discussing the approach further, Ausloos argues it’s important to move away from “binary thinking” to break the current ‘data access’ trust deadlock. “Rather than this binary thinking of disclosure vs opaqueness/obfuscation, we need a more nuanced and layered approach with varying degrees of data access/transparency,” he says. “Such a layered approach can hinge on types of actors requesting data, and their purposes.”

A market research purpose might only get access to very high level data, he suggests. Whereas medical research by academic institutions could be given more granular access — subject, of course, to strict requirements (such as a research plan, ethical board review approval and so on).

“An independent institution intermediating might be vital in order to facilitate this and generate the necessary trust. We think it is vital that that regulator’s mandate is detached from specific policy agendas,” says Ausloos. “It should be focused on being a transparency/disclosure facilitator — creating the necessary technical and legal environment for data exchange. This can then be used by media/competition/data protection/etc authorities for their potential enforcement actions.”

Ausloos says many discussions on setting up an independent regulator for online platforms have proposed too many mandates or competencies — making it impossible to achieve political consensus. Whereas a leaner entity with a narrow transparency/disclosure remit should be able to cut through noisy objections, is the theory.

Advertisement

The infamous example of Cambridge Analytica does certainly loom large over the ‘data for research’ space — aka, the disgraced data company which paid a Cambridge University academic to use an app to harvest and process Facebook user data for political ad targeting. And Facebook has thought nothing of turning this massive platform data misuse scandal into a stick to beat back regulatory proposals aiming to crack open its data troves.

But Cambridge Analytica was a direct consequence of a lack of transparency, accountability and platform oversight. It was also, of course, a massive ethical failure — given that consent for political targeting was not sought from people whose data was acquired. So it doesn’t seem a good argument against regulating access to platform data. On the contrary.

With such ‘blunt instrument’ tech talking points being lobbied into the governance debate by self-interested platform giants, the AlgorithmWatch report brings both welcome nuance and solid suggestions on how to create effective governance structures for modern data giants.

On the layered access point, the report suggests the most granular access to platform data would be the most highly controlled, along the lines of a medical data model. “Granular access can also only be enabled within a closed virtual environment, controlled by an independent body — as is currently done by Findata [Finland’s medical data institution],” notes Ausloos.

Another governance structure discussed in the report — as a case study from which to draw learnings on how to incentivize transparency and thereby enable accountability — is the European Pollutant Release and Transfer Register (E-PRTR). This regulates pollutant emissions reporting across the EU, and results in emissions data being freely available to the public via a dedicated web-platform and as a standalone dataset.

“Credibility is achieved by assuring that the reported data is authentic, transparent and reliable and comparable, because of consistent reporting. Operators are advised to use the best available reporting techniques to achieve these standards of completeness, consistency and credibility,” the report says on the E-PRTR.

Advertisement

“Through this form of transparency, the E-PRTR aims to impose accountability on operators of industrial facilities in Europe towards to the public, NGOs, scientists, politicians, governments and supervisory authorities.”

While EU lawmakers have signalled an intent to place legally binding transparency requirements on platforms — at least in some less contentious areas, such as illegal hate speech, as a means of obtaining accountability on some specific content problems — they have simultaneously set out a sweeping plan to fire up Europe’s digital economy by boosting the reuse of (non-personal) data.

Leveraging industrial data to support R&D and innovation is a key plank of the Commission’s tech-fuelled policy priorities for the next five+ years, as part of an ambitious digital transformation agenda.

This suggests that any regional move to open up platform data is likely to go beyond accountability — given EU lawmakers are pushing for the broader goal of creating a foundational digital support structure to enable research through data reuse. So if privacy-respecting data sharing frameworks can be baked in, a platform governance structure that’s designed to enable regulated data exchange almost by default starts to look very possible within the European context.

“Enabling accountability is important, which we tackle in the pollution case study; but enabling research is at least as important,” argues Ausloos, who does postdoc research at the University of Amsterdam’s Institute for Information Law. “Especially considering these platforms constitute the infrastructure of modern society, we need data disclosure to understand society.”

“When we think about what transparency measures should look like for the DSA we don’t need to reinvent the wheel,” adds Mackenzie Nelson, project lead for AlgorithmWatch’s Governing Platforms Project, in a statement. “The report provides concrete recommendations for how the Commission can design frameworks that safeguard user privacy while still enabling critical research access to dominant platforms’ data.”

Advertisement

You can read the full report here.

TechCrunch

Keep an eye on what we are doing
Be the first to get latest updates and exclusive content straight to your email inbox.
We promise not to spam you. You can unsubscribe at any time.
Invalid email address

NEWS

We asked ChatGPT what will be Google (GOOG) stock price for 2030

Published

on

We asked ChatGPT what will be Google (GOOG) stock price for 2030

Investors who have invested in Alphabet Inc. (NASDAQ: GOOG) stock have reaped significant benefits from the company’s robust financial performance over the last five years. Google’s dominance in the online advertising market has been a key driver of the company’s consistent revenue growth and impressive profit margins.

In addition, Google has expanded its operations into related fields such as cloud computing and artificial intelligence. These areas show great promise as future growth drivers, making them increasingly attractive to investors. Notably, Alphabet’s stock price has been rising due to investor interest in the company’s recent initiatives in the fast-developing field of artificial intelligence (AI), adding generative AI features to Gmail and Google Docs.

However, when it comes to predicting the future pricing of a corporation like Google, there are many factors to consider. With this in mind, Finbold turned to the artificial intelligence tool ChatGPT to suggest a likely pricing range for GOOG stock by 2030. Although the tool was unable to give a definitive price range, it did note the following:

“Over the long term, Google has a track record of strong financial performance and has shown an ability to adapt to changing market conditions. As such, it’s reasonable to expect that Google’s stock price may continue to appreciate over time.”

GOOG stock price prediction

While attempting to estimate the price range of future transactions, it is essential to consider a variety of measures in addition to the AI chat tool, which includes deep learning algorithms and stock market experts.

Finbold collected forecasts provided by CoinPriceForecast, a finance prediction tool that utilizes machine self-learning technology, to anticipate Google stock price by the end of 2030 to compare with ChatGPT’s projection.

According to the most recent long-term estimate, which Finbold obtained on March 20, the price of Google will rise beyond $200 in 2030 and touch $247 by the end of the year, which would indicate a 141% gain from today to the end of the year.

Advertisement
2030 GOOG price prediction: Source: CoinPriceForecast

Google has been assigned a recommendation of ‘strong buy’ by the majority of analysts working on Wall Street for a more near-term time frame. Significantly, 36 analysts of the 48 have recommended a “strong buy,” while seven people have advocated a “buy.” The remaining five analysts had given a ‘hold’ rating.

1679313229 737 We asked ChatGPT what will be Google GOOG stock price
Wall Street GOOG 12-month price prediction: Source: TradingView

The average price projection for Alphabet stock over the last three months has been $125.32; this objective represents a 22.31% upside from its current price. It’s interesting to note that the maximum price forecast for the next year is $160, representing a gain of 56.16% from the stock’s current price of $102.46.

While the outlook for Google stock may be positive, it’s important to keep in mind that some potential challenges and risks could impact its performance, including competition from ChatGPT itself, which could affect Google’s price.


Disclaimer: The content on this site should not be considered investment advice. Investing is speculative. When investing, your capital is at risk.

Source link

Keep an eye on what we are doing
Be the first to get latest updates and exclusive content straight to your email inbox.
We promise not to spam you. You can unsubscribe at any time.
Invalid email address
Continue Reading

NEWS

This Apple Watch app brings ChatGPT to your wrist — here’s why you want it

Published

on

Apple Watch Series 8

ChatGPT feels like it is everywhere at the moment; the AI-powered tool is rapidly starting to feel like internet connected home devices where you are left wondering if your flower pot really needed Bluetooth. However, after hearing about a new Apple Watch app that brings ChatGPT to your favorite wrist computer, I’m actually convinced this one is worth checking out.

The new app is called watchGPT and as I tipped off already, it gives you access to ChatGPT from your Apple Watch. Now the $10,000 question (or more accurately the $3.99 question, as that is the one-time cost of the app) is why having ChatGPT on your wrist is remotely necessary, so let’s dive into what exactly the app can do.

What can watchGPT do?

Source link

Keep an eye on what we are doing
Be the first to get latest updates and exclusive content straight to your email inbox.
We promise not to spam you. You can unsubscribe at any time.
Invalid email address
Continue Reading

NEWS

Discord goes all in with AI: chatbots, automods, whiteboards and more

Published

on

Discord goes all in with AI: chatbots, automods, whiteboards and more

AI is the future, at least over on Discord.

The messaging application originally made for gamers has become Gen Z’s favorite online hangout destination of choice, and now it’s rolling out a number of features powered by artificial intelligence.

In an announcement(Opens in a new tab) on Thursday, Discord shared what’s coming to the platform soon: an AI chatbot, an automated AI moderator, a conversation summarizer, an avatar remixer, and a whiteboard. Some of these features begin rolling out today, March 9. Others will launch in the coming weeks and months.

While AI has jumped into the mainstream thanks to the popularity of OpenAI’s ChatGPT chatbot, Discord has had an active AI community for quite a while now. According to the company, third-party AI apps already on the platform already have more than 30 million monthly users. Nearly 3 million servers on Discord have some AI element integrated into the community.

In fact, the biggest community on Discord is Midjourney, a text-to-image AI project which allows users to generate art from right within the server. Discord says Midjourney’s server has more than 13 million members.

So, with AI being such an integral part of Discord already, it seemed like only a matter of time before Discord itself started bringing AI directly into the platform.

Advertisement

images

AutoMod AI
Credit: Discord

The first feature coming to some Discord servers as soon as today is AutoMod AI. Discord already has an AutoMod feature, which basically automatically moderates rooms for admins based on the rules of the server. Discord has now integrated OpenAI-powered AI into AutoMod, allowing it to search the server and contact moderators when it thinks rules are possibly being broken. According to Discord, AutoMod AI can also consider the context of a conversation so, for example, users don’t get penalized for posts that are misconstrued.

Clyde is a bot that Discord users may already be familiar with, and starting next week, Clyde is getting an AI upgrade. Currently, the Clyde bot provides information, such as server error messages, and also responds to timeout or ban requests from users and mods. However, that’s pretty much all Clyde was able to do. Until now.

Clyde chatbot

Clyde
Credit: Discord

Clyde will now be able to answer all sorts of questions from users, much like OpenAI’s ChatGPT chatbot. Users simply have to type “@Clyde” followed by their prompt. Clyde will be able to pull up information and also help find specific emojis or GIFs based on a user’s description.

Another AI feature coming to Discord next week is Conversation Summaries. Again, the name is fairly descriptive of what it does. With users all over the world, many Discord channels are always moving regardless of time of day. Conversation Summaries will allow users to catch up on what they missed on a Discover Server. The AI-powered feature will “bundle” chats into topics so users can easily read up on what they find most interesting.

Conversation Summaries

Conversation Summaries
Credit: Discord

Starting today, developers can start playing with Avatar Remix, an open-source Discord app that integrates AI art into the messaging app. Avatar Remix allows users to take a fellow user’s avatar and change it up “using the power of generative image models.” What does that mean? In the demo that Discord showed Mashable, a user was able to add a party hat or a mustache to a friend’s avatar by simply mentioning their username and describing what changes they’d like to make.

Advertisement
Avatar Remix

Avatar Remix
Credit: Discord

The company is also launching an “AI incubator,” offering support for developers creating AI-powered apps on Discord.

Finally, Discord revealed a feature that’s coming soon that has long been requested by the Discord community: a whiteboard. But, of course, this won’t be just any collaborative whiteboard feature. It’s going to be AI-powered, allowing users to collaborate in generating AI art and more.

Source link

Keep an eye on what we are doing
Be the first to get latest updates and exclusive content straight to your email inbox.
We promise not to spam you. You can unsubscribe at any time.
Invalid email address
Continue Reading

Trending

Follow by Email
RSS