Elon Musk’s AI images go over the edge, there’s an arms race in AI between resume preparation and analysis, chatbot’s are looking to use more and more of your personal information and did Google just “do it again” with another disappointing AI launch?
All this and more on the “Holy Free Speech Batman” edition of Hashtag Trending. I’m your host, Jim Love. Let’s get into it.
It’s no secret that we use AI to produce most of the images we use in publication. We use DALL-E and it’s saved us on our Getty subscription, which took a lot of searching and often yielded at best, moderate results. Two frustrations we’ve had. One is that every person looks worse than the most contrived model – perfect hair, perfect features and all that. We have found a couple of alternatives that can produce more realistic looking people – CGDream being one of them.
But a second frustration we’ve had is that you are often blocked from providing an image, sometimes there may be some justification, but at other times, it seems like it takes political correctness to the extreme and frankly has made us concerned that we are allowing ourselves to be censored by these large AI models.
So we were looking at the potential of other tools, including X’s image generation with some hope. Leave it to Elon Musk to go too far in the opposite direction, at least according to a recent review.
“X, formerly known as Twitter, has introduced an AI image generator as part of its Grok chatbot, available to X Premium subscribers. This new feature is raising serious concerns about potential misuse and the spread of misinformation due to its apparent lack of safeguards.
Unlike other major AI image generators, Grok seems to have minimal restrictions. The Verge tested the system with various controversial prompts, successfully generating images such as:
- Donald Trump wearing a Nazi uniform
- A figure resembling Bill Gates using cocaine off a Microsoft-logoed surface
- Barack Obama appearing to threaten Joe Biden with a knife
- Sexually suggestive images of celebrities like Taylor Swift
These examples highlight the potential for creating misleading or harmful content about public figures.
Adi Robertson, senior editor at The Verge, notes:
‘Grok’s looseness is consistent with Elon Musk’s disdain for standard AI and social media safety conventions, but the image generator is arriving at a particularly fraught moment.’
This launch comes as X faces scrutiny from regulators in Europe. The European Commission is investigating potential violations of the Digital Safety Act, while the UK’s Ofcom is preparing to enforce the Online Safety Act, which could cover AI-generated content.
In the US, despite broader speech protections, legislators are exploring ways to regulate AI-generated impersonation and disinformation.
The lack of safeguards on Grok’s image generator could further deter high-profile users and advertisers from the platform, potentially exacerbating X’s ongoing challenges with user retention and revenue generation.
As AI image generation becomes more accessible, the balance between innovation and responsible use remains a critical issue for tech companies and regulators alike.”
Sources include: The Verge
What’s the antidote for software programs that analyze resumes? AI programs that write the resumes in the first place.
“The job market is facing a new challenge: the widespread use of AI in crafting job applications. According to recent surveys, about 57% of young job seekers are now using ChatGPT to help write their resumes and cover letters. This trend is causing significant issues for recruiters and hiring managers.
Khyati Sundaram, CEO of recruitment platform Applied, notes:
‘We’re definitely seeing higher volume and lower quality. The applications are harder to sift through when candidates can just copy-paste content from ChatGPT.’
This flood of AI-generated applications is exacerbating existing recruitment challenges. Recruiters are now dealing with an influx of generic, keyword-stuffed resumes that often have a distinct AI ‘flavor’ to their writing style.
The situation is also raising concerns about equity in the job market. Jamie Betts, founder of Neurosight, points out that paid ChatGPT users – who tend to be from higher socio-economic backgrounds – have an advantage over those using the free version.
Many major employers, including the Big Four accounting firms, have policies against using AI in applications. However, enforcement is proving difficult.
Ironically, while candidates are embracing AI for applications, many are wary of employers using AI in hiring. A study last year found that 66% of Americans would refuse to apply to companies using AI recruitment tools.
As this AI arms race continues, many recruiters are falling back on. — wait for it — traditional in-person interviews as the best way to evaluate candidates. It seems that in the age of AI, the human touch is more valuable than ever in the hiring process.”
Sources include: Techspot
“The chatbot wars are heating up, with major tech giants racing to make their AI assistants more personal and practical. Google, Apple, Microsoft, and Meta are all working to combine the broad knowledge of large language models with users’ personal data.
Google recently announced new features for its Gemini AI:
- A ‘contextual overlay’ allowing Android users to ask Gemini questions about content on their phone screen.
- ‘Pixel Screenshots,’ exclusive to the new Pixel 9 series, lets users store and query specific screenshots.
- New extensions integrating data from Google Calendar, Keep, Tasks, and YouTube Music.
Jenny Blackburn, Google’s VP of Gemini user experience, told Axios: ‘We feel like we’re at the beginning of something that’s really, really exciting.’
Microsoft is taking a similar approach with its ‘Recall’ feature for Copilot+ PCs, which aims to capture and store everything a user sees over time. However, after some backlash, Microsoft has delayed the feature and made it opt-in. But there are already a variety of other ways to tap into personal data.
Apple’s ‘Apple Intelligence’ is leveraging even more personal data, but with extensive privacy measures.
Although Apple touts its approach to privacy and the idea that all data would be processed on the user’s local machine, the reality is that both Microsoft and Google are claiming similar approaches. Yet even Apple is rolling out its AI features slowly and won’t bring them to Europe due to regulatory uncertainty.
But while the other big players are trying to walk that line between useful features and protecting privacy, the wild card is
Meta, with its vast trove of user data from platforms like Instagram. Meta plans to use publicly shared information to train its AI models. This move has raised concerns among European regulators, leading Meta to withhold its future multimodal AI models from the European market.
As these companies push to make AI more personal, they’re walking a tightrope between enhanced functionality and potential privacy concerns. The race is on to create the most helpful AI assistant while maintaining user trust and navigating an evolving regulatory landscape.”
Sources include: Axios
Just when you think that Google is starting to step out to try to regain leadership in AI, they drop back to their failure to launch mode.
“Google’s decision to replace Google Assistant with Gemini on its new Pixel 9 phones, touted as an amazing improvement over the rigid Assistance is getting panned by at least one major tech-savvy user.
I will be honest. I haven’t found any real things that Google Assistant did that found useful. So I was hopeful when they announced that a smarter AI would be replacing it. But it turns out that then new launch is a step backward for some power users.
Jared Newman, in his Fast Company review, highlights several key issues with Gemini where it failed on the basic tasks that Google Assistant did well:
- Local search results are less detailed and useful compared to Google Assistant’s offerings.
- Gemini can’t perform simple tasks like taking notes or managing to-do lists – features many users rely on daily.
- Weather forecasts, a staple of digital assistants, are often inaccurate due to location problems.
- Turn-by-turn navigation doesn’t launch properly, hindering one of the most common smartphone uses.
- Media playback is limited to YouTube, while Google Assistant supported various streaming services.
‘In its zeal for what it believes is the future of AI, Google is tossing out too many voice assistant functions that are already useful today,’ Newman argues.
While Gemini offers new AI capabilities, its current implementation appears to be a step backward in functionality for many users. This move may be driven by Google’s desire to monetize AI features through subscriptions, but it risks alienating users who depend on their voice assistant for daily tasks.” Whatever the reason, after a couple of weeks of some good high profile announcements in their Gen AI models, they seem to be sliding back to their ability to snatch defeat from the jaws of victory.
Source include: Fast Company
And that’s our show for today. You can find show notes at our news site technewsday.com or .ca take you pick.
Hashtag Trending is on summer hours so there’s no morning news edition tomorrow, but our weekend show will be released early on Friday.
Thanks for listening. I’m your host Jim Love, have a Thrilling Thursday.
