The weird way AI assistants get their names

cross-posted from: https://feddit.org/post/4455823

Archived link

Kenya is hosting unprecedented lawsuits against Meta Inc., the parent company of Facebook, WhatsApp, and Instagram. Mercy Mutemi, who made last year’s TIME 100 list, is a Nairobi-based lawyer who is leading the cases. She spends her days thinking about what our consumption of digital products should look like in the next 10 years. Will it be extractive and extortionist, or will it be beneficial? What does it look like from an African perspective?

Question: Behind the legal battle with Meta are workers and their conditions. What challenges do they face in these tech roles, particularly content moderation?

Mercy Mutemi: Content moderators in Kenya face horrendous conditions. They’re often misled about the nature of the work, not warned that the work is going to be dangerous for them. There’s no adequate care provided to look after these workers, and they’re not paid well enough. And they’ve created this ecosystem of fear — it’s almost like this special Stockholm syndrome has been created where you know what you’re going through is really bad, but you’re so afraid of the NDA that you just would rather not speak up.

[...]

Content moderation work, annotation work, and algorithm training, [...] in its very nature involves a lot of exposure to harmful content. That work is dumped on Kenya. Kenya says it’s interested in digital development, but what Kenya ends up getting is work that poses serious risks, rather than meaningful investment in its people or infrastructure.

[...]

When the initial version of ChatGPT was released, it had lots of sexual violence in it. So to clean up an algorithm like that, you just teach it all the worst kinds of sexual violence [...] if you ask ChatGPT to show you the worst rape that could ever happen, there are now metrics in place that tell it not to give out this information because it’s been taught to recognize what it’s being asked for. And that’s thanks to Kenyan youth whose mental health is now toast, and whose life has been compromised completely

[...]

Big Tech is not planting any roots in the country [of Kenya] when it comes to hiring people to moderate content or train algorithms for AI. They’re not really investing in the country in the sense that there’s no actual person to hold liable should anything go south. There’s no registered office in Kenya for companies like Meta, TikTok, OpenAI.

[...]

Instead, what you have are these middlemen. They’re called Business Process Outsourcing, or BPOs [...] It’s almost like they’re agents of big tech companies. So they will do big tech’s bidding. If the big tech says jump, then they jump. So we find ourselves in this situation where these companies purely exist for the cover of escaping liability.

[...]

[The workers'] mental health is destroyed – and there are often no measures in place to protect their well-being or respect them as workers. And then it’s their job to figure out how to get out of that rut because they still are a breadwinner in an African context, and they still have to work, right? And in this community where mental health isn’t the most spoken-about thing, how do you explain to your parents that you can’t work?

[...]

I think when you give people work for a period of time and those people can’t work again because their mental health is destroyed, that doesn’t look like lifting people out of poverty to me. That looks like entrenching the problem further because you’ve destroyed not just one person.

[...]

MM: Let me just be very categorical. My position is not that this work shouldn’t be coming into Kenya. But it can’t be the way it is now, where companies get to say “either you take our work and take it as horrible as it is with no care, and we exploit you to our satisfaction, or we, or we leave.” No. You can have dangerous work done in Kenya, but with appropriate level of care, with respect, and upholding the rights of these workers. It’s going to be a long journey to achieve justice.

[...]

[Edit typo.]

cross-posted from: https://feddit.org/post/4456663

Archived link

He Zongying, a female technician employed at a FinDreams Battery factory in Wuhan, a subsidiary of Chinese EV giant BYD, suffered verbal abuse and threats from a male colleague. She immediately reported the issue to the company and also contacted the police. But, to her surprise, the company initially refused to hand over the surveillance footage of the incident to authorities, provided her with no clear path to file a complaint, and then chose to fire her rather than pursue her claim.

Meanwhile, the official union tasked with protecting her rights remained silent. Thus, despite the company’s stated commitment to corporate responsibility, its mechanisms for resolving claims of harassment proved not only ineffective but even resulted in the victim being punished by the company for raising the issue in the first place.

[...]

[Later on], she was called into a meeting with two male colleagues from the factory and a female staff member from Human Resources. But, instead of addressing her complaint, the meeting resulted in He being written up for two violations of company policy. To add insult to injury, one of these violations even stated that she had humiliated and verbally abused the man who had harassed her. She expressed her confusion, pointing out that the incident was on tape and asking why her attempt to stop his harassment was now being characterised as “humiliation and abuse.” In response, her supervisor merely stated that she could have handled the situation differently.

[...]

The second penalty He Zongying received was for exhibiting “improper conduct” and engaging in “repeated misbehaviour within a three-month period.” In addition to the initial altercation, this penalty also involved He allegedly leaving her workstation without authorisation on other occasions.

[...]

Over the following week or so, she was instructed to await further notice and was told that she needed not come into work. When she was finally instructed to return, however, she found that she had been removed from all technical tasks and instead given janitorial duties. Meanwhile, all the days she had been forced to take off by the company had been deducted from her salary.

[...]

BYD Contradicts its own Corporate Responsibility Policies, Union Offers No Support

In its labour rights protection policy, FinDreams Battery states that employees are valuable assets and offers various channels through which employees can provide feedback. The company’s 2023 “Environmental, Social, and Corporate Governance Report” even outlines specific complaint mechanisms and mental health support services offered to workers. The parent firm BYD has also stated in its own corporate responsibility statements that it “firmly opposes workplace violence and harassment.”

[...]

[Edit typo.]

cross-posted from: https://poptalk.scrubbles.tech/post/1628369

Workers are striking after trying a failing multiple times to negotiate fair pay, standardize Remote/Hybrid options, and other basic rights. The strike comes the day before election day.

I guess fingers crossed at the NYT that the election website doesn't tip over from too much traffic....

cross-posted from: https://feddit.org/post/4447520

Seven French families have filed a lawsuit against social media giant TikTok, accusing the platform of exposing their adolescent children to harmful content that led to two of them taking their own lives at 15, their lawyer said on Monday.

The lawsuit alleges TikTok's algorithm exposed the seven teenagers to videos promoting suicide, self-harm and eating disorders, lawyer Laure Boutron-Marmion told broadcaster franceinfo.

The families are taking joint legal action in the Créteil judicial court. Boutron-Marmion said it was the first such grouped case in Europe.

"The parents want TikTok's legal liability to be recognised in court", she said, adding: "This is a commercial company offering a product to consumers who are, in addition, minors. They must, therefore, answer for the product's shortcomings."

TikTok, like other social media platforms, has long faced scrutiny over the policing of content on its app.

cross-posted from: https://feddit.org/post/4454212

Archived link

[...] Thousands of businesses, executives and entrepreneurs [...] rely on the island to turn their AI visions into reality. From Nvidia Corp. and Microsoft Corp. to OpenAI, the world’s AI frontrunners are increasingly turning to Taiwanese companies to fabricate their chips, build their servers and cool their devices. That in turn has made the island’s stock market the hottest major bourse in Asia over the past year, led by Taiwan Semiconductor Manufacturing Co. and Hon Hai Precision Industry Co.

[...]

There are risks for Taiwan. For the first time in decades, an entire technology production ecosystem will be centered not in China but its tiny neighbour. Growing tensions between the US and China may have dissuaded some AI companies from producing hardware in the mainland. Yet the rising importance of Taiwan makes it all the more alluring for Beijing, which has long described the island as a breakaway province it will eventually reclaim.

[...]

The island is now chock-full of lesser-known firms that are just as essential for global AI development. These linchpins include server maker Quanta Computer Inc., power leader Delta Electronics Inc. and Asia Vital Components Co., a pioneer in creating computer cooling systems. Collectively, Taiwanese firms are poised to play an outsized role in an AI market that’s projected to reach $1.3 trillion by 2032.

[...]

As its economy grew, [Taiwanese] companies became more sophisticated manufacturers and began to open factories in mainland China. But they always kept their most advanced techniques at home.

In recent years, the increasingly aggressive US trade sanctions on China have forced companies to scout out alternative production locations, knocking the country out of many supply chains.

In less than two years, for example, those curbs have effectively sidelined China’s AI hardware industry. Taiwan’s exports of servers and graphics cards — the building blocks of data centers for training AI models — in the first nine months of 2024 were more than double China’s output, according to data collected by Bloomberg. That’s a sharp reversal from previous years.

[...]

Archived link

The largest book publisher in The Netherlands has confirmed it plans to use artificial intelligence (AI) to translate some of its books into English, The Bookseller can exclusively reveal.

Utrecht-headquartered publisher Veen Bosch & Keuning (VBK) was acquired by Simon & Schuster earlier this year. It was Simon & Schuster’s first acquisition of a non-English-language publisher, which it said at the time would help it access “broader European markets”.

A spokesperson for VBK told The Bookseller: “We are working on a limited experiment with some Dutch authors, for their books to be translated into English language using AI. There will be one editing phase, and authors have been asked to give permission for this.

[...]

Ian Giles, chair of the Translators Association at the Society of Authors (SoA), said: “This is concerning news. Earlier this year, the SoA found that one third of literary translators are already losing work to AI. Where work itself is not lost, translators struggle to increase their prices in the face of the AI challenger. This pressure on translators’ incomes jeopardises our ability to support ourselves in what is a highly precarious industry.”

[...]

cross-posted from: https://feddit.org/post/4415864

The petitioner calls for the European Union to actively develop and implement a Linux-based operating system, termed ‘EU-Linux’, across public administrations in all EU Member States.

This initiative aims to reduce dependency on Microsoft products, ensuring compliance with the General Data Protection Regulation (GDPR), and promoting transparency, sustainability, and digital sovereignty within the EU.

The petitioner emphasizes the importance of using open-source alternatives to Microsoft 365, such as LibreOffice and Nextcloud, and suggests the adoption of the E/OS mobile operating system for government devices. The petitioner also highlights the potential for job creation in the IT sector through this initiative.

[Edit typo.]

For anyone OOTL with Userbenchmark:

Userbenchmark - the April Fools that never ends

cross-posted from: https://feddit.org/post/4423551

The European Commission choose a consortium to deploy the IRIS² satellite constellation, as the European Union seeks its own satellite broadband system to ensure the bloc isn’t reliant on foreign entities such as Elon Musk’s Starlink.

The SpaceRISE consortium is composed of three European satellite network operators: SES SA, Eutelsat SA, and Hispasat S.A. The consortium relies on a Core Team of European subcontractors from all segments of the satcom ecosystem for the delivery of the scope of the concession contract: Thales Alenia Space, OHB, Airbus Defence and Space, Telespazio, Deutsche Telekom, Orange, Hisdesat and Thales SIX.

The procurement is now entering its final phase towards the signature of the concession agreement scheduled by December 2024, upon which the legal and financial commitment from both parties will be taken.

Some military forces such as those of the U.S. rely on Elon Musk's Starlink, the first large satellite constellation to use low Earth orbit to deliver broadband Internet, but this has been raising concerns of dependency on a private company that could potentially deactivate a critical service.

A Chinese aerospace company has successfully completed the first test flight of a groundbreaking hypersonic passenger aircraft.

Some experts say Google is just parroting your own beliefs right back to you. It may be worsening your own biases and deepening societal divides along the way.

[...]

"Google's whole mission is to give people the information that they want, but sometimes the information that people think they want isn't actually the most useful," says Sarah Presch, digital marketing director at Dragon Metrics, a platform that helps companies tune their websites for better recognition from Google using methods known as "search engine optimisation" or SEO.

[...]

"> What Google has done is they've pulled bits out of the text based on what people are searching for and fed them what they want to read" – Sarah Presch

archive.is link

But WIRED reviewed posts from American Patriots Three Percent (AP3) groups and profiles that are still on the platform, including examples where members and leaders brandish AP3 insignia and share photos from their in-person training sessions.

There have also been some recent instances where Facebook has even auto-generated pages for militias. In May, Facebook auto-generated a page for AP3’s Arizona chapter. In June, Facebook auto-generated a page for “AP3 NM [New Mexico] Training Range.” If you hover over the information widget on the page, Facebook’s explainer reads: “This unofficial page was created because people on Facebook have shown interest in this place or business. It’s not affiliated with or endorsed by anyone associated with AP3 Training Range.” (Meta, Facebook’s parent company, has repeatedly come under fire in the past for auto-generating pages for extremist, white supremacist, and terrorist organizations; a whistleblower first flagged the issue in 2020 in a supplement to an earlier petition filed with the Securities and Exchange Commission.)

Direct link to paper: https://ojs.aaai.org/index.php/AIES/article/view/31748/33915

Ironic, just saw the post about ChatGPT launching their own competitor