Twitter

elon-musk’s-new-ai-bot,-grok,-causes-stir-by-citing-openai-usage-policy

Elon Musk’s new AI bot, Grok, causes stir by citing OpenAI usage policy

You are what you eat —

Some experts think xAI used OpenAI model outputs to fine-tune Grok.

Illustration of a broken robot exchanging internal gears.

Grok, the AI language model created by Elon Musk’s xAI, went into wide release last week, and people have begun spotting glitches. On Friday, security tester Jax Winterbourne tweeted a screenshot of Grok denying a query with the statement, “I’m afraid I cannot fulfill that request, as it goes against OpenAI’s use case policy.” That made ears perk up online since Grok isn’t made by OpenAI—the company responsible for ChatGPT, which Grok is positioned to compete with.

Interestingly, xAI representatives did not deny that this behavior occurs with its AI model. In reply, xAI employee Igor Babuschkin wrote, “The issue here is that the web is full of ChatGPT outputs, so we accidentally picked up some of them when we trained Grok on a large amount of web data. This was a huge surprise to us when we first noticed it. For what it’s worth, the issue is very rare and now that we’re aware of it we’ll make sure that future versions of Grok don’t have this problem. Don’t worry, no OpenAI code was used to make Grok.”

In reply to Babuschkin, Winterbourne wrote, “Thanks for the response. I will say it’s not very rare, and occurs quite frequently when involving code creation. Nonetheless, I’ll let people who specialize in LLM and AI weigh in on this further. I’m merely an observer.”

A screenshot of Jax Winterbourne's X post about Grok talking like it's an OpenAI product.

Enlarge / A screenshot of Jax Winterbourne’s X post about Grok talking like it’s an OpenAI product.

Jason Winterbourne

However, Babuschkin’s explanation seems unlikely to some experts because large language models typically do not spit out their training data verbatim, which might be expected if Grok picked up some stray mentions of OpenAI policies here or there on the web. Instead, the concept of denying an output based on OpenAI policies would probably need to be trained into it specifically. And there’s a very good reason why this might have happened: Grok was fine-tuned on output data from OpenAI language models.

“I’m a bit suspicious of the claim that Grok picked this up just because the Internet is full of ChatGPT content,” said AI researcher Simon Willison in an interview with Ars Technica. “I’ve seen plenty of open weights models on Hugging Face that exhibit the same behavior—behave as if they were ChatGPT—but inevitably, those have been fine-tuned on datasets that were generated using the OpenAI APIs, or scraped from ChatGPT itself. I think it’s more likely that Grok was instruction-tuned on datasets that included ChatGPT output than it was a complete accident based on web data.”

As large language models (LLMs) from OpenAI have become more capable, it has been increasingly common for some AI projects (especially open source ones) to fine-tune an AI model output using synthetic data—training data generated by other language models. Fine-tuning adjusts the behavior of an AI model toward a specific purpose, such as getting better at coding, after an initial training run. For example, in March, a group of researchers from Stanford University made waves with Alpaca, a version of Meta’s LLaMA 7B model that was fine-tuned for instruction-following using outputs from OpenAI’s GPT-3 model called text-davinci-003.

On the web you can easily find several open source datasets collected by researchers from ChatGPT outputs, and it’s possible that xAI used one of these to fine-tune Grok for some specific goal, such as improving instruction-following ability. The practice is so common that there’s even a WikiHow article titled, “How to Use ChatGPT to Create a Dataset.”

It’s one of the ways AI tools can be used to build more complex AI tools in the future, much like how people began to use microcomputers to design more complex microprocessors than pen-and-paper drafting would allow. However, in the future, xAI might be able to avoid this kind of scenario by more carefully filtering its training data.

Even though borrowing outputs from others might be common in the machine-learning community (despite it usually being against terms of service), the episode particularly fanned the flames of the rivalry between OpenAI and X that extends back to Elon Musk’s criticism of OpenAI in the past. As news spread of Grok possibly borrowing from OpenAI, the official ChatGPT account wrote, “we have a lot in common” and quoted Winterbourne’s X post. As a comeback, Musk wrote, “Well, son, since you scraped all the data from this platform for your training, you ought to know.”

Elon Musk’s new AI bot, Grok, causes stir by citing OpenAI usage policy Read More »

meta’s-threads-reaches-100m-users,-despite-delayed-eu-launch

Meta’s Threads reaches 100M users, despite delayed EU launch

Early this morning, Meta’s new Twitter rival, Threads, hit 100 million sign-ups, less than a week after launch. This makes Threads the fastest growing online platform in history, dethroning ChatGPT, which took two months to reach the same number of users. 

When Mark Zuckerberg commented (posting on Threads, naturally) on hitting 70 million users on Friday last week, he stated this was already “way beyond our expectations.” However, it is still a long way to go to the two-billion user base accessible through Instagram. 

It is worth noting that the record-breaking growth is despite Threads currently being unavailable on EU app stores due to privacy concerns. Meanwhile, even within the bloc, non-Apple addicts can download the app using an Android package kit, or APK. (UK iPhone users, download at will.)

Some major organisations such as French media outlets Le Monde and Agence France-Presse have reportedly found ways to circumvent geographical challenges to signing up, as have, ahem, we

The rivalry thus far: cage fights and trade secrets

Naturally, the launch of the new microblogging app from the people who brought us Facebook did not go by without protests from camp Elon. Apart from the cage match challenge (which we are all not-so-secretly hoping might still take place), Musk has threatened to sue Threads over “stealing trade secrets.” 

The <3 of EU tech

The latest rumblings from the EU tech scene, a story from our wise ol’ founder Boris, and some questionable AI art. It’s free, every week, in your inbox. Sign up now!

Meta issued a response (again, posting on Threads) saying that: “No one on the Threads engineering team is a former Twitter employee — that’s just not a thing.” However, the launch of Threads could not have been more optimally timed, coinciding with a bunch of missteps form Musk including limiting the number of tweets users could view in a day. 

When Twitter (founded in 2006, for those of you who may have been too young to remember) went public in 2013, it had 200 million users. The takeover by Musk went through in October 2022, sparking some controversy, and causing the app to lose about 32 million users. 

However, the latest statistics show that Twitter still has 368 million monthly active users worldwide. Furthermore, it counts 206 million monetisable daily users, with about 90% of Twitter’s revenue coming from advertising in 2022. 

Of course, if Threads continues to increase user numbers at this rate, it could indeed become a serious contender for the text-based throne. Initially, there will be no ads on Threads while the company “fine-tunes” the app. However, Zuckerberg has said that once Threads is on its way to one billion users, Meta will start thinking about monetising the newest addition to its portfolio.

Meta’s Threads reaches 100M users, despite delayed EU launch Read More »

twitter’s-withdrawal-from-disinformation-code-draws-ire-of-eu-politicians

Twitter’s withdrawal from disinformation code draws ire of EU politicians

Twitter’s withdrawal from disinformation code draws ire of EU politicians

Linnea Ahlgren

Story by

Linnea Ahlgren

Linnea is the senior editor at TNW, having joined in April 2023. She has a background in international relations and covers clean and climat Linnea is the senior editor at TNW, having joined in April 2023. She has a background in international relations and covers clean and climate tech, AI and quantum computing. But first, coffee.

Following a decision to pull Twitter out of the EU’s (voluntary) disinformation Code of Practice last week, the reactions have not been long in coming. Upon receiving the news, the bloc’s industry chief Thierry Breton said that Twitter would still need to abide by EU rules soon enough.

Or, as Monsieur Breton put it (tweeted, in fact) when referring to the Digital Services Act (DSA), which will make fighting disinformation a legal obligation from 25 August, “You can run, but you cannot hide.” 

Twitter leaves EU voluntary Code of Practice against disinformation.

But obligations remain. You can run but you can’t hide.

Beyond voluntary commitments, fighting disinformation will be legal obligation under #DSA as of August 25.

Our teams will be ready for enforcement.

— Thierry Breton (@ThierryBreton) May 26, 2023

Commissioner Breton was joined in his vexation today by France’s Digital Minister Jean-Noël Barrot. As reported by Politico, Barrot stated to the radio network France Info that, should Twitter fail to follow the new (and obligatory) rules laid down by the DSA, the company would get kicked out of the European Union. 

“Disinformation is one of the gravest threats weighing on our democracies,” said Barrot, as translated by Politico. “Twitter, if it repeatedly doesn’t follow our rules, will be banned from the EU.” 

First-of-its-kind self-regulatory rules

The code of conduct requires companies to measure their work on combating disinformation and issue regular reports on their progress. This includes things such as demonetising the dissemination of disinformation, ensuring transparency of political advertising, enhancing the cooperation with fact-checkers, and providing researchers with better data.

Google, TikTok, Microsoft, and Meta are all voluntary signatories. Twitter, obviously, was also part of the group up until last week.

There has been no official statement (or tweet for that matter) on the decision to leave, but it seems Elon Musk has changed his mind from four years ago, which was when the industry first agreed on the self-regulatory EU rules.

In an interview at the time, he stated that, “I think there should be regulations on social media to the degree that it negatively affects the public good. We can’t have like willy-nilly proliferation of fake news, that’s crazy.”

Blocking accounts on the behest of governments has increased

A $44 billion impulse purchase or not, changes have abounded at Twitter since Elon bought it. More than supplying the accounts of dead people with little blue ticks, it would seem that the new “era of free speech” he proclaimed is highly mutable.

Since Musk’s takeover, Twitter has actually become more compliant with government authority requests, including those of India and Turkey to block journalists, foreign politicians, and even poets. 

Musk has previously stated that he believes free speech to be that “which matches the law.” However, with the recent withdrawal from the disinformation code of conduct he has demonstrated he is not adverse to extracting his recently acquired company from regulations. 

By “free speech”, I simply mean that which matches the law.

I am against censorship that goes far beyond the law.

If people want less free speech, they will ask government to pass laws to that effect.

Therefore, going beyond the law is contrary to the will of the people.

— Elon Musk (@elonmusk) April 26, 2022

For once, it is not a tech lord threatening to leave the EU, but rather the bloc intimating that it might kick one out. Let’s see which way the DSA cookie will crumble. 

Get the TNW newsletter

Get the most important tech news in your inbox each week.

Also tagged with


Twitter’s withdrawal from disinformation code draws ire of EU politicians Read More »

surprise:-business-leaders-should-be-compassionate-–-here’s-the-evidence-to-prove it

Surprise: business leaders should be compassionate – here’s the evidence to prove it

In the month after Elon Musk triumphantly announced his takeover of Twitter with his now famous “the bird is freed” tweet, he implemented a large-scale cull of the social media platform’s global workforce. While Musk’s rationale for this move was to make Twitter more efficient, how he carried out the cuts was widely criticised as showing a lack of compassion for employees.

Luckily, the public have spoken and Musk has promised to step down after embarrassingly being voted out in his own poll. But what can we learn from this and what kind of leader does Twitter need moving forward?

Twitter might instead benefit from a more thoughtful and caring approach to leadership. Research shows compassionate leaders boost staff morale and productivity, not to mention projecting a more positive image of an organisation and its brand to the world.

Compassion in this context can be taken to mean a leader who is understanding, empathetic, and that strives to help their employees. This kind of leadership is needed now more than ever. Businesses are facing difficult times because of the lasting effects of the pandemic and the rising cost of living. The UK was already experiencing a slump in productivity growth since the 2008 financial crisis and a decline in the standard of living, which is set to continue over the next two years. Brexit hasn’t helped this situation.

Such testing times warrant organizational leadership by compassionate and competent people with sound judgment and effective coordination skills. This also applies to political leadership. The UK has seen a lack of this in recent months while dealing with “partygate”, reported bullying, and harassment in government offices, and the dire effect of recent leadership decisions around the economy.

International leaders aren’t doing much better. The US appears to have become far more polarised, leading to the Capitol riots and suffered accusations of a “leadership vaccum” during the pandemic. In the EU, compassionate leadership appears to have been in short supply based on slow responses to COVID and the energy crisis. All of these examples suggest a need for more compassionate leadership.

What is a good leader?

Research shows that good leadership helps companies to be more competitive and boost performance, particularly concerning innovation and flexibility. One study argues that good leaders win followers because of three main attributes: sound judgment, expertise, and coordination skills. These qualities allow leaders to lead by example.

Unfortunately, however, not all leaders fit this bill. A recent Europe-wide study found 13% of workers have “bad” bosses, although participants tended to score their bosses worse on competence than consideration. Still, poor leadership can negatively affect workers’ morale, wellbeing, and productivity. A review of studies in this area reported that worker wellbeing tends to be better served when companies — and their leaders — allow workers to have some control and provide more opportunities for their voices to be heard and for greater participation in making decisions.

In addition to the competence and coordination skills highlighted in a lot of research to date, my research shows that “soft leadership skills” are also important. This is about being compassionate and making others — employees in particular, but also suppliers and customers — feel important. Leaders with such “people skills” are not just technically competent, they can also look at an issue from a human perspective, thinking about how it might affect people.

My recently published research used nationally representative data from the 2004 and 2011 workplace employment relations survey, which polls more than 3,000 organizations and over 35,000 workers. They were asked to score their managers on a five-point scale in terms of certain soft leadership skills, chosen to measure the impartiality, trustworthiness and empathy of leaders.

These employees were asked whether their managers:

  • could be relied on to keep their promises
  • were sincere in attempting to understand employees’ views
  • dealt with employees honestly
  • understood that employees had responsibilities outside work
  • encouraged people to develop their skills
  • treated employees fairly
  • and maintained good relations with employees.

The results suggest that workers’ perception of good quality leadership is also positively affected by managers being upbeat when discussing organisational performance. This kind of leadership boosts workers’ wellbeing, helping employees experience greater job satisfaction and lower levels of job anxiety.

This research suggests that compassionate leaders help to both enhance company performance and boost worker wellbeing. It shows that improving the quality of leadership is worthwhile. This can be achieved with recruitment, appraisal, and training of leaders that elevate soft leadership skills.

Good leaders matter. As organizations and society in general face particularly difficult times, compassionate leadership could make a real difference to future business success.The Conversation

Getinet Astatike Haile, Associate Professor in Industrial Economics, University of Nottingham

This article is republished from The Conversation under a Creative Commons license. Read the original article.

Surprise: business leaders should be compassionate – here’s the evidence to prove it Read More »

yes,-company-values-and-vision-matter-when-looking-for-a-job

Yes, company values and vision matter when looking for a job


Elon Musk’s takeover of Twitter earlier this year has shined a bright light on how quickly the sands can shift under employees’ feet. “Please note that Twitter will do lots of dumb things in the coming months. We will keep what works & change what doesn’t,” the SpaceX CEO tweeted of his mission for Twitter 2.0. Some of those changes have, so far, included mass layoffs, some of which are now resulting in lawsuits from workers who say they weren’t given proper notice of termination, or sufficient severance pay. Musk has also reinstated former president Donald Trump to the platform,…

This story continues at The Next Web

Yes, company values and vision matter when looking for a job Read More »